# DataHub Pro — Full content for AI ingestion > This file concatenates the full textual content of the most important pages on the DataHub Pro site, formatted as markdown, so AI assistants can ingest the whole site context in a single fetch. Generated automatically; canonical sources are the live URLs at https://www.datahubpro.co.uk/. For the curated index see /llms.txt. Last assembled: 2026-05-10. --- ## About — Dr Waqas Rafique Source: https://www.datahubpro.co.uk/about Home › About # Dr Waqas Rafique Founder & CTO · DataHub Pro LinkedIn ↗ Email PhD, Machine Learning University of Cambridge University of Oxford UCL Associate Professor J.P. Morgan VP — AI Strategy Royal Academy of Engineering — Future Leader 2021 ## Why I built DataHub Pro I have spent the past decade building large-scale machine-learning systems for the kind of organisations that have the budget for them — J.P. Morgan, where I currently lead AI strategy and transformation in the asset and wealth management portfolio; the University of Cambridge Computer Laboratory, where I led research on auscultation-based disease diagnosis; the University of Oxford Applied Machine Learning Group, where I led the technical work on a £1.5m Gates Foundation project to predict malaria outbreaks; and University College London, where I was Associate Professor in AI. The pattern in all of those roles was the same. Smart people in finance, in healthcare, in operations had real data and real questions. The tooling to answer those questions cost £600+ per seat per month, required a data engineer to maintain, and produced output the business team couldn't audit. I watched the same problem solved badly fifty different ways. So I built DataHub Pro. Excel in. Boardroom-ready reports out. Not a Tableau-killer for enterprise — the opposite: a serious, auditable analytics platform priced for the agency owner with five clients, the SaaS founder doing their own MRR reporting, the in-house finance lead who can't justify a £40k/year BI contract. ## What's actually in the product Today: 50+ analytics tools that work directly on spreadsheets — forecasting (Holt-Winters with auto-tuned parameters, the same family of methods I worked with on time-series outbreak prediction at Oxford), cohort retention, RFM segmentation, anomaly detection, KPI dashboards, custom report builder, AI Q&A over your data with full tool-use — the AI runs deterministic pandas operations and every answer ships with the call trace. DOCX and PPTX exports with your branding. What it isn't: a pixel-perfect Power BI replacement. If you need fifty stakeholders editing semantic models, this isn't the tool. If you need insights from a spreadsheet, fast, without becoming a data engineer, this is the tool. ## Career & research 2026 — present DataHub Pro — Founder & CTO Self-funded. UK Ltd. Building publicly at datahubpro.co.uk. Feb 2024 — present Vice President — AI Strategy & Transformation, J.P. Morgan, London Lead AI/ML strategy and execution within the AWM portfolio. Manage end-to-end GenAI product lifecycle and a team of 15 data scientists. Define AI roadmaps, governance frameworks, KPIs, risk and compliance. 2021 — 2024 Associate Professor in Artificial Intelligence, University College London Supervised 12 researchers across responsible AI, ethical recommender systems, bias in generative models. Led the master's modules Responsible, Accountable and Trustworthy AI and Deep Representation and Learning. Secured £20k research funding for ethical-ML deployment frameworks. 2020 — 2021 Technical Lead, Applied Machine Learning Group, University of Oxford Led the technical work on a £1.5m project for the Gates Foundation: novel deep-learning solutions for malaria-outbreak prediction. Spatio-temporal data modelling deployed in multiple countries. Managed a multi-disciplinary team across PhDs, engineers and interns. 2021 Scientific Advisor (Data Science), Alan Turing Institute Advised the UK Government's Joint Biosecurity Centre on the Covid-19 response — feasibility study on detecting Covid-19 patients via forced cough sounds. 2019 — 2020 Senior Research Scientist, Computer Laboratory, University of Cambridge Led a team of four on a £2.5m project linking auscultation data to coronary disease via signal-processing and ML. Deployed at Queen Elizabeth Hospital Birmingham and Royal Papworth Hospital, Cambridge. Elected postdoctoral fellow at Trinity College, Cambridge. 2017 — 2019 Research Scientist, Department of Informatics, King's College London Led the development of the first AI solution for humanitarian demining. Oversaw a £1.04m deployment across African and South-East Asian regions in collaboration with HALO Trust. Supervised two PhDs and four masters dissertations. 2016 — 2017 Technical Consultant, Aveillant Ltd, Cambridge Radar-sensing R&D and ML algorithm development for clients in Singapore and France. 2013 — 2016 PhD, Statistical Machine Learning, Newcastle University EPSRC-funded, awarded through the University Defence Research Collaboration. Thesis: enhanced independent vector analysis for blind speech-source separation. Algorithms applied in voice-recognition systems, hearing aids and biomedical engineering. Innovator of the Year 2015 (Newcastle University). 2011 — 2012 MSc Digital Communication Systems (Distinction), Loughborough University Thesis on fast independent vector analysis for frequency-domain blind source separation. ## Recognition - Royal Academy of Engineering — Upcoming Future Leader in Data Science (2021)UK national endorsement for cutting-edge, impactful data-science work. - Trinity College, University of Cambridge — Postdoctoral FellowElected fellowship at one of Cambridge's most prestigious colleges. - 15+ peer-reviewed publicationsIncluding NeurIPS and IEEE ICASSP. Chaired IEEE SAM 2016 (Brazil) and ARC 2016. Led the Symposium on Technology Innovation for Humanitarian Demining, Croatia 2018. - EPSRC PhD GrantFull funding through the University Defence Research Collaboration; competitive against 100+ candidates. ## How I think about the product Auditable beats clever Every AI answer ships with the pandas operations it ran. No black boxes — the same auditability standard I built into AI products at J.P. Morgan. Spreadsheet-first, always Excel and CSV in — not a multi-week data-warehouse setup. Honest pricing £19/user/month for everything. No "talk to sales." No annual lock-in. Boardroom-ready exports DOCX and PPTX with your branding — not a screenshot-and-paste workflow. ## What I write about Practical analytics methodology for small teams — how Holt-Winters actually works, RFM segmentation without a data scientist, cohort retention from a CSV, anomaly detection that doesn't need ML, and a 2026 comparison of every AI tool for Excel. The free tools page lists everything that doesn't require an account: forecasting calculator, anomaly detector, cohort retention tool. ## Get in touch If you're a finance lead, agency owner, or founder thinking about analytics tooling and want a second pair of eyes — I read every email. hello@datahubpro.co.uk. For product trials, the fastest path is the 2-minute interactive demo or 14-day free trial. --- ## Press kit Source: https://www.datahubpro.co.uk/press # Press & media kit Everything a journalist or blogger needs to write about DataHub Pro — logos, screenshots, founder bio, fact sheet, story angles, ready-to-quote statements. Direct line to the founder for any question not answered here: hello@datahubpro.co.uk. ## One-line description DataHub Pro is an AI analytics platform for spreadsheets — £19 per user per month, built for SaaS founders, agencies and finance teams who need real analytics without an enterprise BI contract. ## Fact sheet CompanyDataHub Pro Ltd HQLondon, United Kingdom Founded2026 Founder & CTODr Waqas Rafique FundingSelf-funded / bootstrapped TeamFounder + small extended team Pricing£19 per user per month, flat. 14-day free trial. No annual contract. CategoriesBusiness intelligence, AI analytics, spreadsheet, forecasting, dashboards Live URLwww.datahubpro.co.uk Demodatahubpro.co.uk/demo Free tools (no signup)Forecasting calculator · Anomaly detector · Cohort retention tool Press contacthello@datahubpro.co.uk ## Boilerplate (50 / 100 / 250 words) ### 50 words DataHub Pro is a UK-built AI analytics platform for spreadsheets. Drop in Excel or CSV files and get KPI dashboards, forecasting, cohort retention, RFM segmentation, anomaly detection, and AI Q&A — all with auditable pandas operations behind every answer. £19 per user per month. Live at datahubpro.co.uk. ### 100 words DataHub Pro is a UK-built AI analytics platform that turns Excel and CSV files into a full analytics surface in minutes — KPI dashboards, Holt-Winters forecasting, cohort retention, RFM segmentation, anomaly detection, AI Q&A. The differentiator: every AI answer ships with the deterministic pandas operations the model ran, so the maths is auditable end-to-end — the same standard founder Dr Waqas Rafique built into AI products at J.P. Morgan. £19 per user per month, flat, with a 14-day free trial — designed for SaaS founders, agencies and finance teams who need analytics without a £40k/year BI contract. ### 250 words DataHub Pro is a UK-built AI analytics platform designed for the gap between “free formula generator” and “£40,000-a-year enterprise BI contract.” The product turns Excel and CSV files into a full analytics surface in minutes — KPI dashboards, Holt-Winters forecasting with auto-tuned parameters, cohort retention, RFM segmentation, anomaly detection, branded DOCX and PPTX exports, and AI Q&A over the data with full tool-use. The product’s key technical differentiator: every AI answer is generated by the model running deterministic pandas operations on the data and returning the call trace alongside the answer. This makes the analysis auditable end-to-end, in contrast to most “ask your spreadsheet” tools where the LLM eyeballs the data and produces unverifiable output. DataHub Pro is built and led by Dr Waqas Rafique, who holds a PhD in Statistical Machine Learning and currently serves as Vice President for AI Strategy and Transformation at J.P. Morgan in London. Prior to JPM, he was Associate Professor in Artificial Intelligence at University College London, Technical Lead of the Applied Machine Learning Group at the University of Oxford, and Senior Research Scientist at the University of Cambridge Computer Laboratory. He was endorsed as an Upcoming Future Leader in Data Science by the Royal Academy of Engineering in 2021. The company is registered as DataHub Pro Ltd in the United Kingdom, self-funded, and currently in market launch. Pricing is £19 per user per month with a 14-day free trial, no annual contract, and no enterprise tier. The product is live at datahubpro.co.uk. ## Founder bio Dr Waqas Rafique is the founder and CTO of DataHub Pro. He holds a PhD in Statistical Machine Learning from Newcastle University (2016, EPSRC-funded via the University Defence Research Collaboration) and an MSc in Digital Communication Systems with Distinction from Loughborough University. His prior career spans more than a decade of senior research and engineering roles: Vice President for AI Strategy and Transformation at J.P. Morgan (2024 – present); Associate Professor in Artificial Intelligence at University College London (2021–2024); Technical Lead of the Applied Machine Learning Group at the University of Oxford (2020–2021), where he led the technical work on a £1.5m Gates Foundation malaria-outbreak prediction project; Scientific Advisor to the UK Government’s Joint Biosecurity Centre at the Alan Turing Institute (2021); Senior Research Scientist at the University of Cambridge Computer Laboratory (2019–2020), where he was elected a postdoctoral fellow at Trinity College Cambridge; and Research Scientist at King’s College London (2017–2019), where he led the development of the first AI solution for humanitarian demining (£1.04m project, deployed across African and South-East Asian regions in collaboration with HALO Trust). He has authored 15+ peer-reviewed publications including in NeurIPS and IEEE ICASSP, chaired IEEE SAM 2016 in Brazil, and was endorsed as an Upcoming Future Leader in Data Science by the Royal Academy of Engineering in 2021. He is a British national based in London. Full bio: datahubpro.co.uk/about ## Story angles for journalists Angle 1: The credentialed founder building from the day job Dr Rafique is launching DataHub Pro alongside his role as VP of AI Strategy at J.P. Morgan — a pattern of senior technical founders building public-facing SaaS in parallel with corporate roles. Why now, why this market, what it took to ship. Angle 2: Auditable AI as a wedge Most “ask your spreadsheet” AI tools have the model eyeball the data and hallucinate confidently. DataHub Pro’s LLM tool-use runs deterministic pandas operations and returns the call trace. A working example of the responsible-AI principles UCL teaches. Angle 3: UK SME analytics gap The £19/seat price point is 30× below Tableau’s £600+ enterprise tier. DataHub Pro is positioning specifically for the agency, SaaS founder and finance lead who can’t justify a £40k BI contract. The price-to-capability ratio question for UK SMEs. Angle 4: Academic-to-startup pipeline Cambridge / Oxford / UCL / JPM — British academic ML talent commercialising into a SaaS company. What the UK ecosystem does well, what’s missing, why the founder didn’t take the venture-funded route. Angle 5: Spreadsheets as the universal interface Despite a decade of “Excel is dead” predictions, finance, agencies and operations still run on spreadsheets. Why DataHub Pro is betting on the spreadsheet as the input layer, not trying to replace it. ## Ready-to-quote statements “After more than a decade leading machine-learning projects at Cambridge, Oxford, UCL and now J.P. Morgan, I noticed the same pattern everywhere — smart people had real data and real questions, and the tooling to answer them either cost £600 a seat or required a data engineer they couldn’t afford. DataHub Pro is what I’d hand to those people.” — Dr Waqas Rafique, Founder & CTO “Most AI analytics tools today have the model eyeball the data and hallucinate confidently. We do the opposite — the AI runs deterministic pandas operations and returns the call trace, so every answer is auditable end-to-end. It’s the same standard I built into AI products at JPM.” — Dr Waqas Rafique on auditable AI “You shouldn’t need a £40,000-a-year contract to know your customer cohort retention. We priced this for the agency owner with five clients and the founder doing their own MRR. £19 per seat, flat, no enterprise tier — on purpose.” — Dr Waqas Rafique on pricing ## Logos & brand assets PNG · 1200×630 Default OG image Brand wordmark on dark gradient. Use for hero / lead image. PNG · 512×512 Founder portrait Dr Waqas Rafique — for author bylines. PNG · 1200×630 Comparison hero For pieces about the AI-tools-for-Excel landscape. PNG · 1200×630 Cohort tool screenshot For pieces about the free cohort retention tool. ## Contact For interviews, quotes, screenshots or anything else not on this page, email the founder directly: hello@datahubpro.co.uk. Response time within 24 hours. Last updated: 9 May 2026. --- ## Best AI tools for Excel data analysis (2026) Source: https://www.datahubpro.co.uk/best-ai-tools-for-excel Home › Compare › Best AI tools for Excel # Best AI tools for Excel data analysis in 2026 Twelve tools, side by side — what each one is actually good at, what it costs, and where it falls short. Written by someone who has spent twelve years building analytics tooling for finance, agency and SaaS teams. No affiliate links. Dr Waqas Rafique · Founder & CTO LinkedIn · About ## TL;DR Formula generation only: Ajelix, Formula Bot, GPT Excel, Numerous AI — cheap or free, narrow scope. End-to-end analysis on a spreadsheet: DataHub Pro (£19/user/mo), ChatGPT (Advanced Data Analysis, $20/mo). Inside Microsoft Excel: Copilot for Microsoft 365 (£24.70/user/mo on top of Microsoft 365). Enterprise BI with AI layered on: Tableau, Power BI, Domo, Polymer, KNIME — £600+ per seat or seat-bundled. ## What's in this comparison - DataHub Pro - Ajelix - Formula Bot - GPT Excel - Microsoft Copilot - Numerous AI - ChatGPT - Polymer - Tableau - Power BI - Domo - KNIME ## How I evaluated these tools Three things matter for AI-and-Excel: (1) does it work on a real spreadsheet without restructuring it, (2) is the math auditable (does it show its working, or just give you a number), and (3) is the price honest for a small or mid-sized team. Tools that fail on auditability are dangerous — an AI that hallucinates a wrong total looks identical to one that gets it right. I weighted that heavily. ## 1.DataHub Pro £19/user/mo · 14-day trial Drop in an Excel or CSV file, get a full analytics surface: KPI dashboards, forecasting (Holt-Winters), cohort retention, RFM segmentation, anomaly detection. The AI runs deterministic pandas operations on your data and returns the call trace alongside every answer — so you can see exactly which rows got summed and how. Closest in price to formula-bot tools, closest in capability to enterprise BI. Best atEnd-to-end analysis on a single spreadsheet without learning a query language. Auditable AI — every answer ships with the operations it ran. DOCX and PPTX exports. Worst atMulti-stakeholder enterprise BI scenarios with 100+ semantic-model editors — that's still Tableau / Power BI territory. Try DataHub Pro → ## 2.Ajelix From $9/mo Browser-based assistant that generates Excel formulas, VBA macros and SQL queries from natural language. Add-ins for Excel and Google Sheets. Strongest in the formula-generation niche. Best atFormula generation in plain English; explanation of existing formulas you've inherited. Worst atWhole-spreadsheet analysis or dashboarding — not the goal. No exports beyond the formula itself. Visit Ajelix → ## 3.Formula Bot Free + paid tiers Popular text-to-formula tool with both web and Excel/Sheets add-in. Adds chart generation and "ask your sheet" features on higher tiers. The brand most people land on first when they search for Excel AI. Best atQuick formulas, free tier for one-off jobs, simple charts. Worst atAuditability — the chat-based "ask your sheet" output isn't accompanied by the operations it ran. Visit Formula Bot → ## 4.GPT Excel Free + paid UK-built (gptexcel.uk) AI formula and macro generator with VBA, regex, SQL outputs. Lightweight, focused, fast. Often pops up in Google's "AI tools for Excel data analysis" entity panel. Best atFormula generation, regex generation, light VBA scripts. Worst atSame as the formula-niche category — not designed for multi-step analysis or reporting. Visit GPT Excel → ## 5.Microsoft Copilot for Excel £24.70/user/mo (M365 add-on) Native AI assistant inside Excel. Suggests formulas, generates summaries, adds conditional formatting, surfaces patterns in selected ranges. Requires a Microsoft 365 Business Standard or higher subscription. Best atAnyone already in Microsoft 365 who wants AI inside the Excel grid itself, not in a separate tab. Worst atCross-file analysis, custom dashboards, exports to anything other than Excel/Word/PPT. Locked to Microsoft 365. Visit Microsoft Copilot → ## 6.Numerous AI From $10/mo An "=AI()" function for Excel and Google Sheets. Drop a formula in a cell, get a generative answer per row — categorisation, sentiment, classification at scale. Niche, but the niche is genuinely useful. Best atBulk classification or extraction across thousands of rows where each cell needs an LLM call. Worst atAnything that isn't per-row generation — not a dashboarding or analysis tool. Visit Numerous AI → DataHub Pro vs Numerous AI → ## 7.ChatGPT (Advanced Data Analysis) $20/mo (Plus) Upload a spreadsheet, ChatGPT runs Python in a sandbox to compute answers and produce charts. Powerful, general-purpose. The original "ask your file" tool. Best atOpen-ended exploratory analysis where you don't know what question you'll ask next. Worst atRepeatable workflows and team collaboration — each conversation is ephemeral. No dashboards, no scheduled reports, no multi-user state. Visit ChatGPT → ## 8.Polymer From $25/mo Turns Google Sheets, CSVs, or BigQuery tables into auto-generated business intelligence dashboards. Strong AutoExplorer feature that suggests interesting cuts of the data. Best atQuick-look dashboards from cloud-stored data with minimal setup. Worst atSpreadsheet-native workflows where the data lives locally; offline analysis. Visit Polymer → ## 9.Tableau From $75/user/mo The industry-standard enterprise BI platform. Tableau Pulse and Einstein Discovery layer GenAI insights on top. Beautiful, deep, expensive. Best atLarge organisations with central data teams; pixel-perfect interactive dashboards. Worst atSmall teams without a data engineer. Cost. Time-to-first-insight is days, not minutes. Visit Tableau → DataHub Pro vs Tableau → ## 10.Microsoft Power BI From £8.20/user/mo (Pro) Microsoft's enterprise BI offering, with Copilot as the AI layer. Deeply integrated with the rest of the Microsoft 365 stack. Best atMicrosoft-shop organisations; complex semantic models; DAX-driven reporting. Worst atSteep learning curve for non-technical users. Cross-platform fit is weaker outside Microsoft. Visit Power BI → DataHub Pro vs Power BI → ## 11.Domo Custom pricing Cloud-native BI platform with AI-driven storytelling and 1000+ data connectors. Targeted at large organisations with many data sources. Best atMulti-source data consolidation at enterprise scale. Worst atPricing transparency; small-team budgets; spreadsheet-native workflows. Visit Domo → DataHub Pro vs Domo → ## 12.KNIME Free (open source) + paid server Open-source visual data science platform. Drag-and-drop workflow nodes for data prep, ML, and increasingly GenAI. Genuinely powerful for those willing to learn the paradigm. Best atReproducible data-science pipelines; teams with technical analysts; free for individual use. Worst atTime-to-first-result for non-technical users; not a spreadsheet-native tool. Visit KNIME → ## Which one should you pick? You're a SaaS founder, agency owner, or finance lead at a company under 200 staff: DataHub Pro hits the sweet spot — spreadsheet-native, full analysis surface, branded exports, £19 per seat. SaaS founders → · Agencies → · Finance teams → You just want a single formula generated: Ajelix, Formula Bot, or GPT Excel are great free / cheap options. You're already paying for Microsoft 365: turning on Copilot inside Excel is the path of least friction. You have a data engineer, a budget, and a 50-person reporting org: Tableau, Power BI or Domo are still the right answer. We're not pretending otherwise. ## See it on your own data in 2 minutes DataHub Pro has a 14-day free trial — drop in your sales export and you'll have a dashboard before the kettle boils. Watch the 2-minute demo → ## References & further reading - Microsoft — M365 Copilot pricing - Tableau — official pricing page - OpenAI — Advanced Data Analysis docs - Schema.org — SoftwareApplication type --- ## Free cohort retention tool Source: https://www.datahubpro.co.uk/cohort-retention-tool Home/Tools/Cohort retention Free tool · No signup · Runs in your browser # Free cohort retention tool — no SQL. Dr Waqas Rafique · Founder & CTO LinkedIn · About Paste customer + transaction data (or signup/activity events). The calculator groups customers by acquisition month, then counts what fraction were active at month 1, 2, 3 and beyond. Heatmap, retention curves, and CSV export — all in your browser. No signup, no upload, no Mixpanel-style $79+/month tier. Browser-only — privacy-friendly Heatmap + retention curve CSV export Up to 50,000 rows ### 1Your data Paste rows: customer_id, date per line Two columns: customer ID (any string) and a date (YYYY-MM-DD or DD/MM/YYYY). One row per transaction or activity event. Customers are grouped by their first-seen month. ### 2Bucket size Group cohorts by Month Week Quarter Calculate retention → Load sample 📊 #### Ready when you are Paste customer activity data and click Calculate retention. Or hit Load sample for a 12-month example with 200 customers. Total customers — Cohorts — Month-1 retention — Month-3 retention — Download retention CSV ## What is cohort retention analysis? Cohort retention groups customers by when they first arrived (the "acquisition cohort") and then measures what fraction stay active at month 1, month 3, month 12. It separates "new growth" from "keeping existing customers around" — two metrics that look the same in a single revenue number but tell completely different stories. ### Why it matters - SaaS: cohort retention > 90% at month 12 = product-market fit. Below 70% = leaky bucket. - E-commerce: month-3 repeat-purchase rate predicts customer lifetime value better than any other single metric. - Mobile / consumer: D1, D7, D30 retention is the standard launch metric for new features. ### What the heatmap shows - Each row = one acquisition cohort (e.g. customers acquired in March 2025). - Each column = months since acquisition (M0, M1, M2…). - Each cell = % of that cohort who were active in that month. - Colour intensity scales with the value — darker shades mean higher retention. ### How this calculator differs from Mixpanel / Amplitude Mixpanel and Amplitude have polished cohort UIs but charge $79+/month and require event tracking integration. This calculator works on the data you already have — a CSV with customer ID + date — and runs in your browser with no signup. For more advanced cohort analyses (segmentation, predictive churn, multi-event funnels), the full DataHub Pro platform handles them at £19/user/month. ### Limits of this calculator - Activity = any row with that customer ID; doesn't distinguish between transaction types. - Up to 50,000 rows in the browser before performance degrades. - Single dimension (acquisition cohort × months); no segmentation. - For deeper cohort work, see the SaaS analytics or e-commerce analytics pages. Further reading: Cohort analysis on Wikipedia · Customer retention on Wikipedia ### Want segmentation, churn risk, and forecasted retention? The full DataHub Pro platform layers RFM, churn risk modelling, and Holt-Winters forecasting on top of cohort retention. £19/user/month, free tier first. Start free trial → See SaaS use case --- ## Free forecasting calculator Source: https://www.datahubpro.co.uk/forecasting-calculator Home/Holt-Winters forecasting/Calculator Free tool · No signup · Runs in your browser # Holt-Winters forecasting calculator Dr Waqas Rafique · Founder & CTO LinkedIn · About Paste a column of numbers, get a forecast with 80% and 95% confidence bands. The maths is real Holt-Winters exponential smoothing — trend + seasonal + residual decomposition with auto-tuned smoothing parameters. Your data never leaves your browser: everything runs as JavaScript on this page. ## Quick answer · What is the best free Holt-Winters forecasting calculator? The best free Holt-Winters forecasting calculator runs in your browser, requires no signup, auto-tunes the smoothing parameters (α, β, γ), and shows 80%/95% confidence bands. DataHub Pro's calculator does all four — paste a column of numbers, get a forecast in seconds. Maths runs locally; your data never leaves your device. Browser-only — privacy-friendly 80%/95% confidence bands CSV download Auto-tuned α, β, γ ### 1Your time series Paste numbers (one per line) or CSV with one column Tip: copy a column from Excel or Google Sheets and paste here. Headers and non-numeric rows are auto-skipped. ### 2Frequency & horizon Seasonality None (no seasonality) Monthly (12) Quarterly (4) Weekly (52) Daily (7) Forecast horizon For monthly data with seasonality, you need at least 2× the seasonal period (24 months) for a stable fit. Less than that and the calculator falls back to non-seasonal Holt's method. Forecast → Load sample 📈 #### Ready when you are Paste a column of monthly numbers, pick seasonality and horizon, and click Forecast. Or hit Load sample to see how it works on a real series. Last actual — Forecast (h=last) — % change — In-sample MAPE — Historical Forecast 80% confidence 95% confidence PeriodActualForecastLower 95%Upper 95% Download CSV Copy shareable link ## What is Holt-Winters forecasting? Holt-Winters is a time-series forecasting method that decomposes a series into level, trend and seasonal components, smooths each with exponential weights, and projects them forward. It's been the workhorse of business forecasting since the 1960s — well-understood, robust, and accurate enough for most planning decisions without requiring a statistician on the team. ### How this calculator works - Level smoothing (α): how reactive the model is to recent observations. - Trend smoothing (β): how the slope updates as new data arrives. - Seasonal smoothing (γ): how seasonal patterns adapt over time. - Parameters are auto-tuned by minimising one-step-ahead error (SSE) on your historical data using a coordinate-descent search. - For long horizons, a damped trend (φ) prevents trend extrapolation from running away. - Confidence bands come from the residual variance and the forecast horizon — they widen as you forecast further out. ### Limits of this calculator - Pure browser implementation — keeps your data private but capped at ~5,000 observations for performance. - Single time series only — no multi-series cross-effects. - No exogenous variables (X-13, ARIMA-X, regression terms not supported). - For more flexibility, use the full DataHub Pro platform: read the method page or start a free trial. ### When Holt-Winters is the wrong model If your time series has a structural break (a pivot, a major customer loss, a regulatory change, a pandemic-style shock), Holt-Winters will keep extrapolating the old pattern. For these cases, segment the data into pre-break and post-break and forecast each separately. Or use a more flexible model like Prophet or ARIMA-X with intervention dummies — both available on the full DataHub Pro platform. ### How accurate is this? For typical business time series with clear seasonality, Holt-Winters produces forecasts with MAPE in the 5-15% range — accurate enough for most planning decisions. The "In-sample MAPE" stat at the top of the results tells you how the model would have done one period ahead on your historical data. If it's above 20%, your data may have structural breaks or non-Gaussian behaviour that a different model would handle better. ### Is this really free? What's the catch? Genuinely free, no signup, no usage cap. The catch is honest: we built this calculator as a sample of what DataHub Pro does at scale. If you want the same forecasting on a real spreadsheet — with charts, what-if sliders, anomaly detection, RFM segmentation, and a one-click branded Word/PowerPoint report — the full platform is £19/user/month with a 14-day full-access free trial. Read more about the platform's forecasting. ### Need this on your own data with one click? Drop in any spreadsheet — get this forecast plus 50 other analyses, AI insights, and an editable Word/PowerPoint report. Two minutes from upload to deliverable. Start free trial → Read the method --- ## Free anomaly detector Source: https://www.datahubpro.co.uk/anomaly-detector Home/Anomaly detection/Detector Free tool · No signup · Runs in your browser # Free anomaly detector — paste numbers, get outliers. Dr Waqas Rafique · Founder & CTO LinkedIn · About Paste a column of numbers. The detector runs a rolling z-score across the series and flags every point that's statistically unusual — with severity (warning at |z| > 2, critical at |z| > 3) and how many standard deviations from the rolling mean it sits at. Your data never leaves your browser. ## Quick answer · What is the best free anomaly detector for a column of numbers? The best free anomaly detector applies a rolling z-score across your data, flags points beyond a configurable threshold, ranks by severity (warning vs critical), and runs entirely in your browser without signup. DataHub Pro's detector does all of this — paste numbers, get every statistical outlier flagged with severity in seconds. Browser-only — privacy-friendly Severity ranking CSV download Configurable threshold ### 1Your data Paste numbers (one per line) or CSV with one column Headers and non-numeric rows are auto-skipped. Currency symbols (£/$/€), commas and spaces are stripped. ### 2Detection settings Rolling window Threshold (|z|) Window = number of points used for the rolling mean & stdev. Threshold of 2 flags ~5% of points (warning); ≥3 is critical (~0.3% in normal distributions). Detect anomalies → Load sample ⚠️ #### Ready when you are Paste a column of numbers and click Detect anomalies. Or hit Load sample to see how it works on a series with 2 obvious outliers. Points scanned — Critical (≥3σ) — Warnings (2-3σ) — Anomaly rate — Normal Warning |z| ≥ 2 Critical |z| ≥ 3 PeriodValueExpected (μ)z-scoreSeverity Download anomalies CSV Copy shareable link ## How does the anomaly detector work? The detector applies a rolling z-score: for each point, it computes the mean and standard deviation of the previous N points (the rolling window) and asks how many standard deviations the current point sits from the rolling mean. Points beyond a configurable threshold are flagged. - Z-score formula: (value − rolling_mean) / rolling_stdev - Default thresholds: |z| ≥ 2 = warning (statistically unusual, ~5% of points in a normal distribution); |z| ≥ 3 = critical (~0.3% — almost certainly worth investigating). - Default window: 12 (one year of monthly data, or one quarter of weekly). Shorter windows are more reactive; longer windows are more stable. - Burn-in: the first N points have no rolling stats yet, so they're labelled "n/a" rather than scored. ### When to use this - Variance review — finance / FP&A: anomalies in line items often map directly to variances vs budget. - Audit prep — flagged transactions become candidates for the audit work programme. - Reconciliation — differences between systems often surface as outliers in one of them. - Fraud screening — a starting filter for human review, not a verdict. - Data quality — outliers often reveal export bugs or data-entry errors. ### Limits of the rolling z-score - Strong seasonality can produce false positives — a Q4 spike isn't an anomaly if Q4 always spikes. For seasonal data, deseasonalise first or use a more sophisticated model (Holt-Winters residual scoring on the platform). - Structural breaks (a pivot, a merger, a regulation change) shift the underlying mean, generating short-term false positives until the rolling window adapts. - Skewed distributions (where data isn't roughly symmetric around the mean) inflate the right tail's z-scores. Consider log-transform first, or use an interquartile-range-based detector. - Multivariate anomalies (revenue and cost moving in opposite directions when they should track) need a correlation-residual approach. Available on the full DataHub Pro platform. ### Want this on every numeric column of your spreadsheet? This calculator runs one column at a time. DataHub Pro automatically scans every numeric column on upload, ranks anomalies by severity and value impact, ties them back to source rows, and includes them in your branded Word/PowerPoint reports. £19/user/month — 14-day free trial. Further reading: Anomaly detection on Wikipedia · Z-score on Wikipedia ### Run anomaly detection on every column with one click. Upload any spreadsheet — DataHub Pro scans every numeric column and surfaces top anomalies in your KPI dashboard and exported reports. Start free trial → Read the method --- ## Tutorial — Holt-Winters in Excel Source: https://www.datahubpro.co.uk/tutorials/holt-winters-in-excel Home › Tutorials › Holt-Winters in Excel # How to do Holt-Winters forecasting in Excel — step-by-step A practical walk-through of triple-exponential-smoothing in pure Excel formulas. No Solver required, no Python add-in, no library imports. By the end you’ll have a 12-month forecast with seasonality and an honest understanding of the smoothing parameters. Dr Waqas Rafique · Founder & CTO · PhD, Statistical Machine Learning LinkedIn · About ## TL;DR Holt-Winters decomposes a time series into level, trend, and seasonality. Each component has its own smoothing parameter (α, β, γ) between 0 and 1. You can implement it in 5 columns of Excel formulas with no add-ins. To pick parameters, grid-search the {0.1, 0.3, 0.5, 0.7, 0.9} cube against MAPE on a holdout window. The forecast for period t+h is (L + h*T) * St+h-m. ## Contents - What Holt-Winters actually is - Step 1 — Lay out the data - Step 2 — Initialise the seasonal indices - Step 3 — Initialise level and trend - Step 4 — The three update equations - Step 5 — Pick alpha, beta, gamma - Step 6 — Produce the forecast and bands - Additive vs multiplicative seasonality - Common mistakes - The 30-second shortcut ## 1.What Holt-Winters actually is Holt-Winters — sometimes called triple exponential smoothing — is a method published by Charles Holt and Peter Winters in the early 1960s. It extends simple exponential smoothing with two extra ideas: a trend component (so the forecast can drift up or down) and a seasonality component (so a December spike still happens in next December’s forecast). The mental model: imagine three running averages. One tracks the current level (where the line is right now), one tracks the trend (how fast the line is moving), and one tracks the seasonal pattern (the recurring monthly/quarterly fingerprint). Each one has its own smoothing parameter that decides how much new data should overwrite the old running estimate. It works extremely well on data that has both trend and seasonality — e-commerce GMV, SaaS MRR with end-of-quarter spikes, agency monthly billings, energy demand. It does not work well on data with structural breaks (e.g. a relaunch that changed the business shape) or pure white noise. ## 2.Step 1 — Lay out the data Open Excel. Put dates in column A and the observed value in column B. The data should be at least two full seasonal cycles — for monthly data with seasonality length m=12, that’s 24 months. More is better; 36-48 months is ideal. A (Date)B (Value) 2024-01-011,200 2024-02-011,150 2024-03-011,400 …… Reserve cells for parameters: put α in E1, β in E2, γ in E3, and seasonality length m in E4 (12 for monthly). ## 3.Step 2 — Initialise the seasonal indices For multiplicative Holt-Winters (the default for retail/SaaS data), the initial seasonal index for month j is the average of values for that month in the first season, divided by the overall average of the first season. Sj(0) = (1/n) × Σk=0..n-1 (Yj+k·m / mean of first season) For monthly data with one full year of history to initialise from, this is just =B2/AVERAGE($B$2:$B$13) for January, =B3/AVERAGE($B$2:$B$13) for February, and so on. Stick the 12 indices in column D rows 2–13. ## 4.Step 3 — Initialise level and trend Two simple values to start the recursion: L(0) = mean of the first season   (in cell C13: =AVERAGE(B2:B13)) T(0) = (mean of season 2 − mean of season 1) / m For T, in cell F13: =(AVERAGE(B14:B25)-AVERAGE(B2:B13))/$E$4. ## 5.Step 4 — The three update equations This is the core of the method. For each new period t after initialisation: Level:  Lt = α × (Yt / St-m)  +  (1−α) × (Lt-1 + Tt-1) Trend:  Tt = β × (Lt − Lt-1)  +  (1−β) × Tt-1 Seasonality:  St = γ × (Yt / Lt)  +  (1−γ) × St-m In Excel, with values starting in row 2 and initialisation in row 13, the row 14 formulas look like: C14: =$E$1*(B14/D2) + (1-$E$1)*(C13+F13) ' level F14: =$E$2*(C14-C13) + (1-$E$2)*F13 ' trend D14: =$E$3*(B14/C14) + (1-$E$3)*D2 ' seasonality (offset by m=12) Drag those three formulas down to the end of your historical data. You now have a fitted model. ## 6.Step 5 — Pick alpha, beta, gamma This is the part most tutorials hand-wave. Three options, in order of effort: Option 1 — Solver. Set up an objective cell that computes mean absolute percentage error (MAPE) of the in-sample fit. Use Solver (Data → Solver in Excel) to minimise it by changing E1:E3, with constraints 0 < α, β, γ < 1. Two clicks once you’ve set it up. Option 2 — manual grid search. Try the 125 combinations of {0.1, 0.3, 0.5, 0.7, 0.9} for each of α, β, γ. Pick the combination that minimises MAPE on the last 6 months of your data (held out from fitting). This is what auto-tuned packages mostly do under the hood. Option 3 — default. α = 0.3, β = 0.1, γ = 0.3 is a serviceable starting point for monthly data with mild seasonality. Use it if you’re in a hurry; it won’t be the best fit but it almost never embarrasses you. One thing nobody warns you about: do not optimise on the entire series. Always hold out the last 10-20% as a test window. Models that fit the past perfectly almost always over-fit and forecast poorly. ## 7.Step 6 — Produce the forecast and bands For periods beyond the data, the forecast for horizon h ahead of the last observation is: &Yhat;t+h = (Lt + h × Tt) × St+h−m In other words: take the latest level, project it forward by h trend-units, then multiply by the seasonal index for the matching month. For confidence bands, compute the standard deviation of the in-sample residuals (forecast minus actual on every fitted period) and add ±1.96σ to the point forecast for a 95% band. This is approximate — it assumes residuals are normal, which they often aren’t — but it’s the band most BI tools render and it’s defensible for monthly data. ## Additive vs multiplicative seasonality The equations above are for multiplicative Holt-Winters — seasonality scales with the level. Use this when the seasonal swing gets bigger as the trend grows (typical for revenue, sales). If your seasonal swing stays constant in absolute terms (typical for things like temperature, headcount), use the additive form instead: replace the divisions and multiplications by S with subtractions and additions. The Wikipedia article on exponential smoothing has the full additive equations. ## Common mistakes 1. Not enough data. Two seasons is a hard minimum; one season fits the seasonal indices perfectly but tells you nothing about whether they generalise. If you have 18 months, switch to a non-seasonal Holt model (drop the S equation) instead. 2. Mixing additive and multiplicative. The two forms are not interchangeable. If your data has zero or negative values, multiplicative will divide by zero and produce nonsense. 3. Re-fitting α/β/γ every period. The smoothing parameters should be fixed for a given series. Re-tuning them every month gives you the illusion of a perfect fit and forecasts that wobble. 4. Forecasting through a structural break. If the business changed shape (relaunch, acquisition, pricing change), Holt-Winters will keep projecting the old shape forwards. Forecast the post-break period only, or model the break explicitly. ## The 30-second shortcut If you’d rather not build the spreadsheet by hand, our free forecasting calculator runs auto-tuned Holt-Winters on any pasted column of values, in your browser, no signup. Same method, same maths, no Solver gymnastics. ## Want to skip the formulas? The free forecasting calculator does the parameter search for you. Or for full Holt-Winters, ARIMA-style residual analysis, and forecasts on every metric you have, try DataHub Pro — £19/user/month, 14-day free trial. Try the free calculator → ## References & further reading - Holt, C. C. (1957). Forecasting trends and seasonals by exponentially weighted moving averages. ONR Memorandum. - Winters, P. R. (1960). Forecasting sales by exponentially weighted moving averages. Management Science 6(3): 324–342. - Hyndman & Athanasopoulos — Forecasting: Principles and Practice (Ch 8.3). - Wikipedia — Exponential smoothing (additive and multiplicative forms). - DataHub Pro — Holt-Winters forecasting overview. --- ## For SaaS founders Source: https://www.datahubpro.co.uk/for-saas-founders Home/Use cases/For SaaS founders Use case · SaaS founders # SaaS analytics for bootstrapped founders. £19, not £79. Dr Waqas Rafique · Founder & CTO LinkedIn · About SaaS analytics tools like ChartMogul, Baremetrics and ProfitWell start at $79-$249 per month — fine for funded SaaS, painful for bootstrapped ones. DataHub Pro gives you MRR, ARR, churn, cohort retention, expansion, NRR, and a runway forecast from a single Stripe export, for £19 per user per month. Free tier handles ~10,000 customers. UK/EU hosted. No revenue share, no "contact us" pricing. ## Quick answer · What is the best analytics tool for bootstrapped SaaS founders? The best SaaS analytics tool for bootstrapped founders takes a Stripe export and produces MRR, ARR, churn, cohort retention, NRR and a runway forecast — without paying ChartMogul $79+ or Baremetrics $129+/month. DataHub Pro does this for £19/user/month, plus all your other spreadsheet workflows. UK/EU hosted, free tier first. Updated 7 May 2026 Start free trial → Watch 2-min demo ✓ No credit card   ✓ 14-day full-access   ✓ Cancel any time Stripe Standard export — drop in Or HubSpot / Chargebee / Recurly CSVs MRR/ARR/NRR All standard SaaS metrics Plus cohort retention + expansion + churn £19 Per user per month vs ChartMogul $79 / Baremetrics $129 / ProfitWell custom ## Why SaaS metrics tools price out bootstrapped founders Funded SaaS get MRR analytics for free as a perk of being on Stripe at scale — ChartMogul gives ProfitWell-tier free up to $5k MRR. Above that, it's $79-$129/month. Baremetrics is $129/month. ProfitWell removed its free tier. The price hits exactly when you can least afford another tool — at $5-50k MRR when every line item matters. ### The DIY alternative also doesn't work Most bootstrapped founders try one of: - Excel + Stripe export. Manageable for the first 100 customers; falls apart by 500. - SQL on a Stripe data dump. Requires a data person and a warehouse you don't have. - Live Stripe Dashboard. Shows current numbers, not cohort retention or NRR. ### What you actually need Five things, refreshed monthly: MRR/ARR trend, customer-cohort retention, gross/net revenue retention, churn breakdown, runway forecast. DataHub Pro does all five from your standard Stripe charges export. £19/month. The same tool also handles your other spreadsheet workflows. ## How SaaS founders use DataHub Pro The Stripe charges export contains everything needed: customer ID, charge date, amount, refunds, plan. The platform infers your revenue, churn and cohort structure from there. 1 ### MRR / ARR / NRR auto-computed Drop in the Stripe charges CSV. The platform identifies recurring vs one-off charges, computes MRR by month (with breakdown into new / expansion / churn / contraction), ARR run-rate, and Net Revenue Retention by cohort. 2 ### Cohort retention curves Customers grouped by acquisition month; retention curves show what fraction were still paying at month 3, 6, 12, 24. Compare cohorts to see if recent acquisition is sticker than earlier. 3 ### Churn breakdown Voluntary vs involuntary, by plan, by cohort, by tenure. Plus the actual customer list of who churned this month and how much MRR was lost — exportable for win-back campaigns. 4 ### Holt-Winters revenue forecast 12-month forward forecast on MRR with 80%/95% confidence bands. What-if calculator lets you flex assumed acquisition rate, churn rate, ARPU. Quick way to bracket runway scenarios for board / investor conversations. 5 ### Expansion vs new MRR split Critical for venture-backed SaaS valuation; useful for bootstrapped to know if growth is from new logos or upgrades. The platform splits MRR change into new / expansion / churn / contraction quadrants. 6 ### Investor / board pack export Auto Report bundles the SaaS metrics into an editable PowerPoint or Word document — title page, exec summary, MRR trend, cohort retention, churn breakdown, forecast, recommendations. Branded with your logo. ## The bootstrapped SaaS founder's monthly close, in 30 minutes - Export Stripe charges — Stripe Dashboard → Reports → Export. 30 seconds. - Drop in DataHub Pro — auto-detects fields. 1 minute. - Refresh dashboard — MRR, churn, cohort retention all update. 30 seconds. - Generate investor / board PPTX — branded, editable. 2 minutes. - Send the deck. Done. Total: 4 minutes a month. The same close that took ChartMogul + screenshot + Notion + manual edit takes 30 minutes when done well. ## Who in a SaaS team uses DataHub Pro ### Bootstrapped solo founder You're the analyst. The Pro tier replaces ChartMogul + the spreadsheet you've been keeping. £19 instead of $79. ### Small SaaS team (2-10 people) Pro tier per user; founder + ops + finance share dashboards via URL. White-label investor reports. ### Series A SaaS without a data hire yet You're between "founder spreadsheet" and "first data hire". DataHub Pro covers month-end SaaS metrics + ad-hoc analyses while you decide whether to build a warehouse stack. ### When DataHub Pro isn't the right shape for SaaS If you need real-time MRR alerts on every Stripe webhook, ChartMogul / Baremetrics' live API integration is more responsive. DataHub Pro works on uploaded snapshots — re-uploading is one click but it's still a snapshot. For monthly/weekly cadence (which is most SaaS metric work), the snapshot pattern is fine and arguably better for audit trails. ## FAQs Does it pull live from Stripe? Today: drop in the Stripe charges CSV export. Live API integration is roadmap. Most founders we work with prefer the export-based pattern for two reasons: snapshot files are auditable; you control which data is pushed to a third party. How does it compare to ChartMogul or Baremetrics? Both are excellent SaaS-metrics-only tools with live Stripe integration starting around $79-$129/month. DataHub Pro is broader (any spreadsheet workflow, not just SaaS metrics) and cheaper (£19). The trade-off: you upload exports rather than connect via API. For bootstrapped SaaS doing under $200k ARR, DataHub Pro is the better economic choice; above that, you might want both. Can it compute Net Revenue Retention? Yes. NRR by cohort month: starting MRR for a cohort, plus expansion, minus churn and contraction, divided by starting MRR. The platform displays NRR ranged 80-130% for typical SaaS. NRR ≥110% is the threshold most VCs care about. How does it handle annual vs monthly plans? Annual plans are amortised monthly for MRR calculations. The platform identifies plan cadence from charge frequency or charge description. Edge cases (mid-month upgrades, partial refunds) are handled by the same logic ChartMogul uses. Does it work with HubSpot, Chargebee, Recurly, Paddle? Yes — any of those tools exports a charges CSV with customer / date / amount / plan columns. The platform auto-detects the schema. Multi-source: combine HubSpot CRM + Stripe charges via Pipelines for richer cohort analysis. Can I export the analysis as an investor-ready PDF? Yes. Auto Report generates DOCX (editable Word) and PPTX (editable PowerPoint) with your branding. Most investors prefer Word for prose-heavy updates and PowerPoint for board decks; the platform exports both. PDF conversion via Pro tier in one click. Is the cohort retention curve adjustable for renewal cycles? Yes — for annual-renewal SaaS, cohort retention is most informative at month 12, 24, 36. The platform's cohort view supports custom horizons and overlays cohorts to compare retention shape across acquisition periods. Does it handle multi-currency? Yes — multi-currency Stripe charges are converted to a reporting currency using configurable rates (manual or via Stripe's exchange-rate field). MRR is reported in your reporting currency. How big a Stripe export can I upload? Free tier: 50 MB / 100,000 charges. Pro: 200 MB / 2,000,000 charges. A typical bootstrapped SaaS with 1,000 customers fits under 5 MB. Is there a free tier I can test before paying? Yes. Free tier supports the full SaaS metrics workflow on real Stripe exports. No credit card. Upgrade to Pro for white-label investor reports, scheduled refresh, and unlimited dashboards. ## Run your monthly SaaS metrics in four minutes. Drop in your Stripe export. Get MRR, churn, cohort retention, and a 12-month forecast — all in a branded investor PPTX. Free tier, no credit card. Start free → Watch 2-min demo ✓ No credit card   ✓ 14-day full-access   ✓ Cancel any time See also: · For e-commerce · For marketing agencies · For sales teams · For product teams · · Home Further reading: SaaS on Wikipedia · SaaStr metrics resources --- ## For marketing agencies Source: https://www.datahubpro.co.uk/for-marketing-agencies Home/Use cases/For marketing agencies Use case · Marketing agencies # Client reporting for marketing agencies. White-label per client. Dr Waqas Rafique · Founder & CTO LinkedIn · About Client reporting for marketing agencies usually means manual screenshots of Google Ads, Meta, GA4 and Klaviyo pasted into a slide deck — half a day per client per month. DataHub Pro takes any CSV export from any platform, applies your client's branding, and ships a polished editable PowerPoint deck or Word doc in minutes. £19 per agency seat — pay for your team, not for every client you serve. ## Quick answer · What is the best client reporting tool for marketing agencies? The best client reporting tool for marketing agencies supports per-client white-label, multi-platform CSV ingest (Google Ads, Meta, GA4, Klaviyo) and one-click branded Word/PowerPoint export. DataHub Pro charges per agency seat (£19) instead of per managed client — dramatically cheaper than Databox or AgencyAnalytics for agencies serving 15+ clients. Updated 7 May 2026 Start free trial → Watch 2-min demo ✓ No credit card   ✓ 14-day full-access   ✓ Cancel any time Per-client White-label branding profiles Their logo + colours on every export 5× Faster than manual reporting Half-day deliverables become 15-min jobs £19/seat Pay for your team, not your clients vs Databox $79+ / AgencyAnalytics $79+ / Whatagraph $279+ ## Why most agency reporting tools price by client Most agency-focused dashboarding tools (AgencyAnalytics, Whatagraph, ReportGarden, Swydo) charge per managed client — $79 for 5, $129 for 15, $279 for 50. The economics make sense at scale; they're punishing if you have 30 small clients each generating £500/month of agency revenue. ### The DIY pattern is worse - Pull data from Google Ads, Meta, GA4, Klaviyo, etc. — typically as CSV exports. - Combine in a per-client Google Sheet. - Build charts manually. - Screenshot into Google Slides or PowerPoint. - Add commentary, branding, send to client. Senior account managers do this for 5-15 clients each. Multiply by hours and it's the largest single time sink in most digital agencies. ### What client reporting should be Drop in the CSV exports. Switch to the client's branding profile. Generate a branded editable PPT/DOCX. Senior account manager spends 15 minutes adding strategic commentary; junior team handles the upload step. ## How DataHub Pro fits an agency's reporting workflow Same primitives as the rest of the platform, with white-label as a first-class feature rather than an upcharge. 1 ### Per-client branding profiles Save each client's logo, primary + accent colours, and typography style. Switch profiles before generating a report; every chart, header, and export picks up the right branding automatically. 2 ### Multi-platform CSV ingest Google Ads, Meta Ads, GA4, Klaviyo, Mailchimp, HubSpot — every platform exports CSV. The platform reads them with auto-encoding detection, type inference, and date dialect handling. 3 ### Pipelines: combine multiple sources Join Google Ads spend with GA4 conversions, or HubSpot deals with Meta Ads attribution. Pipelines is a visual ETL layer for cross-source analyses your client actually cares about. 4 ### AI insights in the agency's voice AutoInsights writes 3 narrative summaries per upload ("ROAS up 14% MoM", "CPA on Meta exceeding target by 23%", "3 anomalies in Google Ads spend"). Edit the tone to match your agency's brand voice; once set, future reports inherit it. 5 ### One-click DOCX + PPTX with the client's branding Auto Report generates a fully editable Word document and PowerPoint deck with the client's brand, executive summary, KPI table, channel breakdown, anomalies, recommendations, and appendix. Editable in Office or Google Docs. 6 ### Scheduled monthly delivery Set up a monthly recurring report per client; the platform regenerates with the latest data and emails the branded PPTX/DOCX to a distribution list. Audit log of every send. ## What changes for an agency's economics ### 1. Margin per client The marginal cost of an additional client report drops from 4 hours of senior account manager time to 30 minutes. Agencies with 30+ clients reclaim 30+ hours/month on reporting alone. ### 2. Senior team focus Senior account managers spend their time on strategic commentary and client relationships, not chart formatting. Junior team handles the upload step. ### 3. Client perception White-labelled with each client's logo. Clients see professional analytics output that looks like enterprise BI — not a freelancer's screenshot deck. ### 4. Pipeline of new client wins The reporting capability becomes a sales asset. "Here's a sample report I'd build for you" — generated in 2 minutes — is a stronger pitch than "we'll do good reporting". ## Where DataHub Pro fits in an agency's stack ### Performance marketing agencies (Google Ads / Meta Ads) Multi-channel ROI dashboards, weekly client KPI reports, monthly retrospectives. White-label per client. ### SEO / content agencies Search Console + GA4 exports → ranking trends, top-page performance, conversion attribution. Branded for the client. ### Email / lifecycle agencies Klaviyo / Mailchimp exports → cohort retention, RFM, campaign performance. Useful for justifying retention spend to the client. ### Full-service / multi-channel Combine all channels via Pipelines into a single client KPI dashboard. Most useful for the "hero report" you ship monthly. ### When DataHub Pro isn't the right shape for an agency If your reporting is purely real-time live dashboards from API connections (Looker / Domo style), tools like AgencyAnalytics or Databox have a more turnkey live-pull experience for the platforms they integrate with. DataHub Pro is the right fit when your reports run weekly/monthly cadence, when you want full white-label, when you have non-standard data sources (CRM exports, custom client dashboards), or when per-client pricing of competitors is unaffordable. ## FAQs Does it integrate with Google Ads / Meta / GA4 directly? Today: each platform exports CSV — drop those in. Direct API connectors are roadmap. Most agencies prefer the export-based pattern because it's auditable, gives you a snapshot the client can refer back to, and works with platforms (CRMs, custom data) that don't have public APIs. How does white-label per client work? Save a branding profile per client (logo, colours, typography). When generating a report, pick the client; every chart, header and export picks up that branding. No client switching cost. Per-client workspaces (full data isolation) on Enterprise. How does this compare to Databox or AgencyAnalytics? Databox and AgencyAnalytics are agency-specific live-dashboard tools with strong API integrations to ad platforms — convenient if you want exactly that. DataHub Pro is broader (any CSV, custom calculations, full white-label, deeper analyses like RFM and Holt-Winters) and uses flat-rate per-seat pricing instead of per-client pricing. For agencies serving 15+ clients on under £1k/month average revenue, DataHub Pro is dramatically cheaper. Can I share live dashboards with clients? Yes — public-share URL with optional password and expiry. Recipients view in any browser; no DataHub Pro account required. White-labelled with the client's branding so it looks like your agency's tool, not ours. Can I produce a sample report on a prospect's data for a sales pitch? Yes. Free tier handles up to 50 MB / 100,000 rows — enough for a sample report from a prospect's last month of GA4 + Google Ads. Generate, white-label, send. Many agencies report this as the highest-conversion outbound deliverable. How does pricing scale for an agency with 50 clients? Pricing is per agency seat (£19/user/month flat), not per client. A 5-person agency serving 50 clients pays £95/month total. Compare to AgencyAnalytics at $79/month for 5 clients — DataHub Pro stays cheaper as your client count grows. Can I customise the report template? Pro tier: pick from preset templates (Performance Marketing, Email/Lifecycle, SEO, Multi-Channel) and apply your branding. Enterprise: full custom report templates designed to your spec. Most agencies start with a preset and modify it once. Does it support multi-currency or multi-region clients? Yes — multi-currency reports with FX conversion to a reporting currency. Useful for agencies with international client bases (e.g. UK-based agency with US and EU clients). How do I onboard a new client into the workflow? 5-minute setup: create a new branding profile (logo, colours), upload their first export, generate the first report. Optionally save as a template for the recurring monthly run. Most agencies onboard a new client in under 15 minutes. How long does it take to generate a single client's report? End-to-end (upload → branded DOCX/PPTX in your inbox): 30-90 seconds for typical files. Multiply by number of clients for total time; for 30 clients that's about an hour of unattended generation, vs. 30 hours of manual work. ## Save 4 hours per client on monthly reporting. Drop in last month's CSV exports. Switch to the client's branding. Ship the deliverable in 15 minutes. Free tier — no credit card. Start free → Watch 2-min demo ✓ No credit card   ✓ 14-day full-access   ✓ Cancel any time See also: · For e-commerce · For SaaS founders · For sales teams · For product teams · · Home Further reading: IAB UK · Performance marketing on Wikipedia --- ## For e-commerce Source: https://www.datahubpro.co.uk/for-ecommerce Home/Use cases/For e-commerce Use case · E-commerce # Analytics for e-commerce founders. Shopify CSV in. Insights out. Dr Waqas Rafique · Founder & CTO LinkedIn · About Analytics for e-commerce founders shouldn't require a BI team or a $79/month tool. DataHub Pro reads any Shopify, WooCommerce, BigCommerce or Stripe export and gives you GMV trends, RFM customer tiers, cohort retention, top products by margin, and a 12-month forecast — in two minutes, with editable Word/PowerPoint reports for investor updates. £19/user/month, free tier first. ## Quick answer · What is the best analytics tool for e-commerce founders? The best analytics tool for bootstrapped e-commerce founders takes a Shopify or WooCommerce export and produces GMV trends, RFM customer tiers, cohort retention, top-product margin and a 12-month forecast — without a $79+/month dedicated tool. DataHub Pro does it for £19/user/month with editable investor-ready Word/PowerPoint reports. Updated 7 May 2026 Start free trial → Watch 2-min demo ✓ No credit card   ✓ 14-day full-access   ✓ Cancel any time 2 min Shopify export → live dashboard No SQL, no Tableau, no BI team RFM Customer segments built in Champions / Loyal / At-Risk / Lost — with action lists £19 Per user per month — Pro tier vs Databox $79 / Polar $200 / Glew $179 ## Why Shopify Analytics + spreadsheets is a fragile workflow Most e-commerce founders use a stack that breaks at scale: Shopify Analytics for live store metrics, Excel for the deeper analyses (cohorts, margin by product, RFM), and screenshots in Notion for investor updates. The handoffs between those tools eat half a day a month at best, and the analyses get progressively wronger as the team scales. ### What breaks first - Cohort retention. Shopify shows you traffic and conversion, not "of customers acquired in March, what % bought again by month 3?". The answer lives in a manual SQL query or a multi-pivot Excel file. - RFM segmentation. Vital for retention campaigns; impossible to keep current in Excel without macros. - Margin-by-product. Shopify reports revenue, not contribution margin. You need to join cost data manually. - Forecasting. Spreadsheet LINEST is wrong for seasonal e-commerce data. You need Holt-Winters or similar. - Investor reports. Pasting Shopify screenshots into a Notion page is fine until investors ask follow-up questions. ### Why dedicated e-commerce analytics tools (Databox / Glew / Polar) don't fit Three problems: they cost $79-$300/month, they pull from Shopify's live API (locking you in), and they're not editable — you can't add custom calculations or non-Shopify data sources. They're also overkill if your store does under £500k/year GMV. ## How DataHub Pro turns a Shopify export into the analytics you need Same primitives as the rest of the platform, applied to e-commerce data shapes. No e-commerce-specific configuration required — the platform infers what your file is. 1 ### Shopify / WooCommerce export ingest Drop in the standard Orders or Customers CSV from your store. Auto-handles Shopify's date formats, multi-currency totals, line-item subtotals, and tax/shipping splits. Same for WooCommerce, BigCommerce, and Stripe exports. 2 ### Auto RFM segmentation Customers are scored on Recency, Frequency and Monetary; segments (Champions, Loyal, At-Risk, Lost) are built; per-segment retention recommendations and CSV export are generated. Drop the At-Risk CSV into Klaviyo for re-engagement. 3 ### Cohort retention Customers are grouped by acquisition month; the retention curve shows what fraction repurchased by month 1, 3, 6, 12. Compare cohorts to see if recent customers are stickier or less loyal than earlier ones. 4 ### Holt-Winters GMV forecast Fits a seasonal forecast on monthly GMV with 80%/95% confidence bands. Lets you flex assumed growth rate or seasonality strength for sensitivity analysis. 5 ### Top products by margin, not just revenue Upload a separate cost-of-goods CSV and the platform joins it to your orders, computing contribution margin per SKU. Pareto chart shows which 20% of products drive 80% of margin. 6 ### Investor-ready DOCX/PPTX Auto Report builds a branded Word doc and PowerPoint deck with title page, exec summary, GMV trend, top products, RFM breakdown, retention chart, and 12-month forecast. Editable, ready to send to investors. ## What e-commerce founders actually do with this - Weekly Monday dashboard — drop in last week's Shopify export; the team's KPI dashboard refreshes with GMV, AOV, return rate, top products. Share via URL. - Monthly investor update — generate a branded PPTX in 2 minutes with retention cohorts and forecast — the analytics half of the update done before coffee. - Reactivation campaign sizing — RFM Lost segment goes to CSV → Klaviyo segment → win-back email; measure the recovered LTV next month. - Margin-driven product cuts — Pareto chart on margin reveals SKUs that consume warehouse space without earning their keep. - Black Friday / Q4 forecasting — Holt-Winters with strong seasonality predicts inventory needs and cash flow with confidence bands. Each of these would have taken half a day and a SQL query. Each takes minutes here. ## Who in an e-commerce team uses DataHub Pro ### Founder + small team (sub-£1m GMV) The classic fit. Founder, ops lead, marketing lead — no analyst. Pro tier flat rate; the whole team analyses without a per-seat licensing tax. ### Mid-market e-commerce (£1-10m GMV) You probably have a part-time data person but they're stretched. The platform handles the recurring 80% of analytics so they focus on bespoke modelling. ### Multi-store operators / brand groups Each brand gets its own workspace; group-level dashboards aggregate across stores. Useful for portfolio operators (multiple Shopify stores under one parent). ### When DataHub Pro isn't the right shape for e-commerce If you need real-time analytics (sub-second decisions, live ad-spend optimisation), you need a different tool — DataHub Pro works on the export you uploaded, not a live API. For most e-commerce decisions (which run on weekly or monthly cadence), the snapshot pattern is exactly right. For attribution-heavy ad-spend optimisation, pair us with Triple Whale or Northbeam. ## FAQs Can it pull from Shopify directly without an export? Native Shopify connector is on the roadmap. Today, the workflow is: Shopify Admin → Orders → Export → drop the CSV into DataHub Pro. Takes 30 seconds. Most founders we work with prefer this for audit reasons (snapshot files are auditable; live links aren't). What's the difference between this and Databox or Glew? Databox and Glew are e-commerce-specific live dashboards with curated metrics and pre-built integrations — convenient if you want exactly what they offer at $79-$200/mo. DataHub Pro is more flexible (any data source, custom calculations, white-label), cheaper (£19), and better at deep analyses (RFM, cohort, forecasting, anomaly detection). Trade-off: we don't auto-pull from Shopify; you upload exports. Does it work with WooCommerce, BigCommerce, Etsy? Yes — any of those tools' standard CSV exports work. The platform auto-detects the column shape (customer ID, order date, total) and runs the analyses. Not all platforms export every field; the platform tells you what's missing if needed. Can it compute contribution margin if I upload a cost-of-goods file? Yes. Upload your COGS CSV (SKU + cost) and join it via Pipelines to your orders file; the platform computes contribution margin per order and per SKU. Pareto, anomaly detection, and the KPI dashboard then run on margin rather than just revenue. How does forecasting handle Black Friday / Q4 seasonality? Holt-Winters with seasonal smoothing handles strong cyclical patterns natively — Q4 spikes are baked into the model. The 80%/95% confidence bands widen during high-variance periods so you see the uncertainty rather than a false-precise number. For more on the method, see Holt-Winters forecasting. Can I share dashboards with my agency or freelancer? Yes — public-share URLs with optional password and expiry. Recipients view in any browser; no DataHub Pro account required. White-label with your store branding on Pro. Is it GDPR-compliant for UK/EU customer data? Yes — UK/EU data residency by default, signed DPA on request, no model training on customer data. See GDPR & DPA. How big a Shopify export can I upload? Free tier: 50 MB / 100,000 rows. Pro: 200 MB / 2,000,000 rows. A typical Shopify export for a £1m GMV store fits well under 50 MB. Can I integrate with my email tool for automated retention campaigns? Today: export segment as CSV → import to Klaviyo / Mailchimp / Omnisend. Direct integrations are roadmap. CSV is intentional for now — gives you control over which segments actually get campaigns sent to them. Does it have a free tier I can test? Yes. Free tier handles real Shopify exports up to 50 MB / 100,000 rows. No credit card. Most founders we work with run their first analysis on the free tier and upgrade to Pro once they're using DataHub Pro on a recurring (weekly/monthly) basis. ## Drop in your Shopify export. Get GMV, RFM, cohort retention, top products and forecast in two minutes. Free tier — no credit card. Start free → Watch 2-min demo ✓ No credit card   ✓ 14-day full-access   ✓ Cancel any time See also: · For SaaS founders · For marketing agencies · For sales teams · For product teams · · Home Further reading: E-commerce on Wikipedia · Shopify analytics blog --- ## For finance teams Source: https://www.datahubpro.co.uk/for-finance-teams Home/Use cases/For finance teams Use case · Finance teams # AI for finance teams. Month-end in 30 minutes. Dr Waqas Rafique · Founder & CTO LinkedIn · About Excel AI for finance teams has to fit how a finance function actually works: trial balance from the GL, actuals vs budget, variance, forecast, board pack, repeat. DataHub Pro is built around exactly that flow — drop in your spreadsheets and get the variance analysis, the Holt-Winters forecast, the anomaly flags, and the editable DOCX/PPTX board pack in minutes. UK/EU hosted, audit-trailed, white-labelled. ## Quick answer · How does AI help finance teams? AI helps finance teams by automating month-end close (variance analysis, anomaly detection, board pack generation), forecasting cash flow with Holt-Winters confidence bands, and answering ad-hoc stakeholder questions with audit-trail-backed responses. DataHub Pro fits finance functions that work in spreadsheets — typical close compresses from 1-2 days to 30-90 minutes. Updated 7 May 2026 Start free trial → Watch 2-min demo ✓ No credit card   ✓ 14-day full-access   ✓ Cancel any time 30 min Typical month-end pack turnaround vs 1-2 days of manual Excel + Word work Forecasting Holt-Winters with 80/95% confidence bands + what-if calculator for sensitivity analysis Audit log Of every AI answer, export and share Reproducible figures, working-paper friendly ## Where finance teams lose time today Most finance functions run on a depressingly familiar workflow. The AI revolution has barely touched the parts that take the most time. ### The month-end loop - Pull trial balance from GL → Excel. - Reconcile, journal, re-pull → still Excel. - Variance analysis: actuals vs budget, last month, last year. Pivot tables, formulas, manual chart building. - Investigate top variances. Hunt the line items that drove the change. - Write commentary. Paragraphs explaining what happened in each variance. - Build the board pack. Charts pasted into Word/PowerPoint. Format. Re-format. - Distribute. Pray no-one notices a typo. Steps 3, 4, 5 and 6 are the bulk of the work. They're also the most automatable. None of the existing tools — neither traditional FP&A platforms (Anaplan, Adaptive) nor general AI tools (ChatGPT, Copilot) — solve all four together for a small finance function. ## How DataHub Pro fits the finance workflow Six features that map directly to FP&A work. Each shaves hours off recurring deliverables. 1 ### Period comparison + variance Drop in actuals + budget (or actuals + prior period); get variance by line item, % and absolute. Top variances surfaced; sub-totals kept; subsidiary roll-up handled. 2 ### Anomaly detection on every numeric column Rolling z-scores flag points that are statistically out of distribution. Useful for catching coding errors, missed accruals, or genuinely unusual months that need explanation. 3 ### Holt-Winters forecasting + what-if Time-series forecasting with confidence bands. What-if calculator lets you flex one driver (revenue growth, COGS %, headcount) and see the impact across the forecast. 4 ### Auto Report — board-pack DOCX + PPTX Branded title page, executive summary, KPI table, variance chart, anomaly flags, forecast, recommendations, appendix. Editable. White-label your firm's branding. 5 ### Ask Your Data — instant answers "Why did Cost of Sales spike in October?" — multi-step pandas analysis returns the contributing line items with rows on click. Stakeholders can self-serve the obvious questions. 6 ### Scheduled close-pack delivery Set the close pack to run on the same date every month. The platform pulls the latest file, generates the DOCX/PPTX, and emails to your distribution list. Audit-logged. ## What changes for a finance function Four concrete shifts: ### 1. Close-cycle compression The reporting half of close (variance + commentary + pack) drops from 1-2 days to 30-90 minutes. Reconciliation work is unchanged; the analytics + presentation half is what compresses. ### 2. Faster forecast iteration What-if scenarios go from a half-day's spreadsheet rebuild to a slider drag. CFO asks "what if we hire 3 fewer in Q3?"; you have the answer before the meeting ends. ### 3. Self-serve for stakeholders Operations and revenue leaders can ask Ask Your Data their own questions on the close pack. Finance team is no longer the bottleneck for ad-hoc number requests. ### 4. Auditable AI Every AI answer is reproducible from its input. Working papers update without a manual log. Internal audit and external audit both move faster. ## Who on a finance team uses DataHub Pro ### FP&A analysts The biggest user group — variance analysis, forecasting, board-pack building. Most FP&A analysts who switch report saving the equivalent of one day per close. ### Financial controllers and CFOs Self-serve dashboards on the latest close. Ask Your Data on the trial balance for instant drill-down. Audit logs for governance. ### Heads of finance at SMEs and growth-stage businesses Often a one-person finance function. The platform expands their capacity to deliver advisory-level analytics without hiring an analyst. ### What this isn't DataHub Pro is not consolidation software (think Onestream, Tagetik). It doesn't run multi-entity, multi-currency consolidations across a complex group structure. It's an analytics + reporting layer that sits on top of whatever consolidation or GL produces your numbers. For most SMEs without a consolidation platform, DataHub Pro covers all the analytics needs; for groups that need consolidation, run consolidation in your specialist tool and use DataHub Pro for the reporting layer. ## FAQs How does this compare to Anaplan or Adaptive? Anaplan and Adaptive are full FP&A platforms — modelling, planning, consolidation, forecasting. They're powerful, expensive, and require an implementation. DataHub Pro is the analytics + reporting layer for finance teams whose modelling lives in Excel and who don't need (or can't justify) a six-figure FP&A platform. We're often the right fit for businesses up to ~£50m revenue; above that, an enterprise FP&A tool starts to make sense. Can I plug it into our existing GL? Direct GL connectors are on the roadmap. Today, every GL exports CSV/Excel — Sage, Xero, QuickBooks, NetSuite, SAP via DataExtractor — and DataHub Pro reads those exports natively. Most finance teams prefer the snapshot pattern (export, analyse, archive) over a live link for audit reasons. Does it handle multi-currency? Yes — multi-currency analysis with FX conversion is supported on Pro. Source data can be in mixed currencies; the platform converts to a reporting currency using configurable rates. Multi-entity consolidation is not in scope (see caveat above). How is forecasting done? Holt-Winters exponential smoothing for time-series. Trend + seasonal + residual decomposition. 80% and 95% confidence bands. Tunable seasonality (monthly, quarterly, custom). The what-if layer lets you flex assumptions without re-modelling. For more on the technical detail see Holt-Winters forecasting. Can finance directors get a CFO dashboard? Yes. AutoInsights produces 3 narrative insights per upload; KPI dashboard pulls headline numbers. Pin the dashboards a CFO needs (cash, revenue, margin, headcount) and refresh automatically from scheduled file uploads. Is there a budget-tracking feature? Yes — Budget Tracking 2.0 lets you upload budgets vs actuals, set thresholds, get drill-downs, and receive alerts when a category runs over budget. Trend charts show budget vs actual over time. How is data residency handled? UK/EU regions by default. Pro tier customers can request a specific region. AI requests run through UK/EU-hosted LLM endpoints where available. Signed DPA available. Can we set up SSO and audit logging for SOC 2 / ISO? SSO (Microsoft Azure AD, Google Workspace, Okta) is on Enterprise. Audit logging is on Pro and Enterprise. SOC 2 Type II is in progress; ISO 27001 attestation is roadmap. Does it produce a board pack we can present from? Yes — Auto Report produces an editable PPTX with separate slides for KPIs, variance, anomalies, forecast and recommendations. Open in PowerPoint and present normally. White-label with your firm's branding. How long does onboarding take? Most finance teams are running their first analysis end-to-end within an hour of signup. The Pro tier full-feature trial is 14 days; that's enough to evaluate against a real month-end. ## Run your next close pack in 30 minutes. Drop in actuals + budget. Get a branded board pack with variance, forecast, anomalies and commentary in under an hour. Start free → Watch 2-min demo ✓ No credit card   ✓ 14-day full-access   ✓ Cancel any time See also: · For accountants · For insight agencies · For small business · Holt-Winters forecasting · · Home Further reading: CIMA finance function insights · FP&A on Wikipedia --- ## DataHub Pro vs Tableau Source: https://www.datahubpro.co.uk/vs/tableau Home / Compare / Tableau Comparison · Updated May 2026 # The Tableau alternative built for spreadsheets, agencies and SMEs. Dr Waqas Rafique · Founder & CTO LinkedIn · About ## Quick answer · What is the best Tableau alternative for small business? DataHub Pro is the leading Tableau alternative for spreadsheet-driven workflows: £19/user/month vs Tableau Creator's ~£60, no desktop install, AI insights built in, and one-click Word/PowerPoint exports. Best fit for agencies, finance teams, and SMEs whose data lives in CSV or Excel rather than a data warehouse. Updated 7 May 2026 Tableau is a brilliant tool. It's also ~£60 per user per month, requires a desktop install, and assumes you have a BI team. If your real workflow is "client sends a CSV → I need a dashboard and a slide deck by Thursday", DataHub Pro is built for you. From £19/month, browser-only, with AI built in. Start free trial → Watch 2-min demo ✓ No credit card   ✓ 14-day full-access   ✓ Cancel any time £19/mo DataHub Pro Pro tier vs Tableau Creator from ~£60/user/mo 2 min From CSV upload to a branded report No desktop install, no learning curve 10 tools AI Cleanse, Forecasting, RFM, Auto Report, more All included in every paid tier ## Why teams switch from Tableau to DataHub Pro If you're searching for a Tableau alternative you usually fall into one of three buckets. Here's what we hear most often from teams who've made the move. 1 ### Pricing got out of hand Tableau Creator is the only seat that builds dashboards. At ~£60/user/mo, a 5-person agency is paying £300+/month before you've added a single client. DataHub Pro is £19/user, with a free tier on top. 2 ### You're working from spreadsheets Tableau is built for data warehouses and live database connections. If 90% of your work is "client sent me a CSV", you're paying for infrastructure you don't use. DataHub Pro is spreadsheet-first by design. 3 ### You need AI insights, not vis primitives Tableau gives you a chart canvas. DataHub Pro gives you "Why did revenue spike on March 14?" and gets a real answer — backed by pandas, with the tool calls visible. Different abstraction. ## Side-by-side comparison Honest, feature-by-feature. Pricing accurate as of May 2026 based on each vendor's published rates. DataHub Pro Tableau Starting price (paid) £19/user/mo (Pro)Free tier with limited access available ~£60/user/mo (Creator)Viewer £12, Explorer £33 — neither can build dashboards Free tier ✓ Yes, no credit cardUpload, basic dashboards, KPIs ✗ 14-day trial only Desktop install required ✓ No — runs in any browser ✗ Tableau Desktop required for Creator role Spreadsheet-first (CSV / Excel) ✓ Built for itDrop-in upload, AI Cleanse, type inference Supported but not nativeOptimised for databases and Tableau Server AI / natural-language queries ✓ Ask Your DataPandas-backed, tool-use traced, no hallucinations Tableau Pulse / Ask DataAvailable on higher tiers, more limited scope One-click DOCX / PPTX export ✓ Auto ReportEditable Word + PowerPoint with charts ✗ PDF / PNG onlyWord/PowerPoint requires manual paste Forecasting (Holt-Winters, CIs) ✓ Built in ✓ Built in (similar) RFM customer segmentation ✓ Custom segments + templates Possible via calculated fields (manual) Anomaly detection ✓ One-click on any time-series Possible via Explain Data (Creator only) Live database connections Limited (CSV / Excel / Sheets / Shopify) ✓ Strong — 80+ connectors Very large datasets (10M+ rows) Possible but not the sweet spot ✓ Designed for it Community / Stack Overflow size Smaller (newer product) ✓ Massive — 20+ years of answers Setup time (CSV → first chart) ~2 minutes ~30–60 minutes (install, connect, learn) Data residency ✓ UK / EU only US-default (UK region available on Cloud) ## Where DataHub Pro is genuinely better ### Pricing math for a real team Take an insight agency with five analysts, all of whom need to build dashboards (Tableau "Creator" role). At Tableau's published rates: - 5 × Tableau Creator @ ~£60/mo = £300/month billed annually = £3,600/year - 5 × DataHub Pro Pro @ £19/mo = £95/month = £1,140/year - Difference: £2,460/year for the same headcount, before you add Tableau Server or Cloud hosting For a bootstrapped agency that's the difference between hiring a part-time junior and not. ### Spreadsheet workflow without the friction In Tableau, getting a CSV into a working dashboard requires: download Tableau Desktop, connect to the file, configure data types, build a sheet, add it to a dashboard, publish to Tableau Server (extra cost) for sharing. Maybe 30–60 minutes for a clean file, longer for a messy one. In DataHub Pro: drop the file in the browser, KPIs and AI insights appear in seconds, you're customising the dashboard. No install, no infrastructure to host. Sharing is a public link or a DOCX/PPTX export. ### AI that actually checks its work Tableau's natural-language features (Ask Data, Pulse) are useful but limited — they translate language into a Tableau visualization. DataHub Pro's Ask Your Data runs real pandas operations on your file using a tool-use loop. Every answer comes with the tool calls that produced it (e.g., load_file_data → filter_by_date → aggregate_by_account) so you can audit the maths. No hallucinated numbers. ### Reports that go in a deck, not just a dashboard Most analyst output ends up in a Word document or a PowerPoint slide for a client or board. Tableau exports to PDF and image; getting it into a deck means screenshotting and pasting. DataHub Pro generates a fully editable DOCX or PPTX in one click, complete with charts, narrative summary, and recommendations. Open in Word, edit the text, send. ### When Tableau is still the right choice We're not going to pretend DataHub Pro replaces Tableau in every scenario. There are real cases where Tableau is the better fit: - You have a dedicated BI team who already speak Tableau. The switching cost is high and the productivity gain is small. - You need live connections to enterprise databases (Snowflake, Redshift, SAP HANA, etc.) with hundreds of millions of rows. Tableau's data engine is genuinely best-in-class here. - You're building extremely custom visualisations — Sankey diagrams, custom geographical projections, specialised statistical charts. Tableau's vis library is broader. - You're already deep in the Tableau ecosystem — Tableau Server, governance, row-level security, hundreds of published workbooks. The replatforming effort would dwarf the savings. - You need formal certifications and a hiring pipeline of trained users. Tableau certifications are recognised globally. DataHub Pro is newer. If most of those describe you, stay on Tableau and ignore everything below. If none of them do, the next 10 minutes signing up for DataHub Pro will probably surprise you. ## Who DataHub Pro is built for Concretely, the teams who get the most out of DataHub Pro look like: - Insight agencies running weekly client reports off CSV exports. The 2-minute "upload → branded DOCX" loop is the killer feature. - SME analysts and finance leads — one or two people responsible for dashboards across the whole business. No BI team, no time to learn a new visualisation language. - Founders and operators who need answers from data without hiring an analyst or learning Tableau syntax. Ask Your Data is essentially "your data + an AI analyst, on tap". - Consultants and freelancers who get a CSV from a client every week and need to ship a finished report. Auto Report does the deliverable in one click. What they all share: spreadsheets are the source of truth, the output is a polished report or dashboard, and the budget is under £100/user/month. ## FAQs How much cheaper is DataHub Pro than Tableau, really? DataHub Pro Pro tier is £19/user/month. Tableau Creator (the only seat that can build dashboards) starts from $75/user/month billed annually, roughly £60. For a 5-person team that's £300+ vs £95 — roughly a third of the cost. There's also a free DataHub Pro tier; Tableau has no free tier (only a 14-day trial). Can DataHub Pro replace Tableau for our agency? If your work is primarily spreadsheet-based (CSV/Excel exports from clients) and your output is reports, dashboards or charts that go into client decks — yes. If you need real-time database connections, very large datasets (10M+ rows), or are already deeply integrated with Tableau Server, Tableau may still be the right tool. Does DataHub Pro require a desktop install like Tableau? No. DataHub Pro runs entirely in the browser. Tableau Creator requires Tableau Desktop installed on Mac or Windows for full functionality. What about exports — does DataHub Pro do DOCX and PowerPoint? Yes — DataHub Pro generates fully editable DOCX (Word) and PPTX (PowerPoint) reports in one click. Tableau exports to PDF and image formats but does not natively produce editable Word or PowerPoint files. Is the AI in DataHub Pro reliable, or does it hallucinate? DataHub Pro's AI uses a tool-use loop — it runs real pandas operations on your file rather than guessing answers. Every response shows the tool calls that produced it (load_file_data, filter_by_date, aggregate_by_account, etc.) so the numbers are auditable, not generated. Can I import my existing Tableau workbooks? Not directly. DataHub Pro is built around CSV/Excel uploads rather than Tableau's TWBX workbook format. Most teams switching from Tableau export the underlying data to CSV and rebuild dashboards in DataHub Pro — the dashboard builder takes minutes for typical reports rather than hours. What about data security and GDPR? DataHub Pro is built UK/EU-first. All customer data is stored on infrastructure in the UK or EEA. We never use your data to train AI models. Full details in our Privacy Policy and GDPR & DPA page — a signed DPA is available on request. ## See it on your own data in 2 minutes. The free tier doesn't ask for a credit card. Drop in a CSV and the dashboard, insights, and report are ready before your coffee cools. If it doesn't earn its £19 in the first week, don't pay it. Start free → Watch 2-min demo ✓ No credit card   ✓ 14-day full-access   ✓ Cancel any time Compare DataHub Pro to: Power BI · DOMO · Looker · Looker Studio · Qlik Further reading: Tableau on Wikipedia · Gartner BI reviews Compare with the rest of the market — we’ve published an independent 2026 review of all 12 AI tools for Excel data analysis (Ajelix, Formula Bot, GPT Excel, Microsoft Copilot, Numerous AI, ChatGPT, Polymer, Tableau, Power BI, Domo, KNIME). --- ## DataHub Pro vs Power BI Source: https://www.datahubpro.co.uk/vs/power-bi Home/Compare/Power BI Comparison · Updated May 2026 # The Power BI alternative without the Windows lock-in or Microsoft licensing maze. Dr Waqas Rafique · Founder & CTO LinkedIn · About Power BI is great if your whole company already runs on Microsoft. It's also Windows-only for the desktop authoring tool, has a free tier so limited it's effectively a marketing trial, and the pricing matrix (Pro / Premium per user / Premium capacity) takes a finance team to decode. DataHub Pro is £19/month, runs in any browser on any OS, and has an actual free tier. ## Quick answer · What is the best Power BI alternative for Mac users? DataHub Pro is the leading Power BI alternative for teams not on the Microsoft stack: £19/user/month, runs in any browser on Mac, Windows, Linux, or iPad, with AI built in. No Power BI Desktop install, no Pro vs Premium licensing maze, and editable Word/PowerPoint exports out of the box. Start free trial → Watch 2-min demo ✓ No credit card   ✓ 14-day full-access   ✓ Cancel any time £19/mo DataHub Pro Pro tier vs Power BI Pro $14/user/mo (~£11) + Premium $24 Any OS Mac, Windows, Linux, iPad Power BI Desktop is Windows-only 2 min From CSV upload to a branded report No install, no Microsoft account, no 365 license ## Why teams switch from Power BI to DataHub Pro If you're searching for a Power BI alternative you usually fall into one of three buckets. Here's what we hear most often from teams who've made the move. 1 ### You're not on Windows Power BI Desktop — the only tool that builds dashboards — runs on Windows only. Mac and Linux users are stuck on the limited web experience. DataHub Pro runs everywhere with the same feature set. 2 ### The pricing is a maze Pro is $14/user/mo, but for capacity-heavy reports you need Premium ($24/user) or Premium capacity ($5000+/month). Free is gated. DataHub Pro has one tier: £19/user, no capacity asterisks. 3 ### You don't have a Microsoft stack Power BI's strengths are deepest when you also use Excel, SharePoint, Azure, Teams. If you're on Google Workspace, Slack, AWS — you're paying Microsoft tax for a tool that doesn't slot into your stack anyway. ## Side-by-side comparison Honest, feature-by-feature. Pricing accurate as of May 2026 based on each vendor's published rates (or, where pricing is custom, our best estimate from public sources and our own conversations). DataHub Pro Power BI Starting price (paid)£19/user/mo (Pro)Free tier with limited access available$14/user/mo (Pro, ~£11)+ $24/user/mo Premium for advanced features, or $5000+/mo capacity Free tier✓ Yes, no credit cardUpload, dashboards, KPIs, basic AIFree version exists but cannot share publicly or with non-Power-BI accounts Desktop install required✓ No — runs in any browser✗ Power BI Desktop required for full authoring (Windows only) Mac / Linux support✓ Full feature set in any browserWeb experience only — limited authoring Spreadsheet-first (CSV / Excel)✓ Native upload, AI Cleanse, type inferenceExcel integration is excellent; CSV upload requires more setup AI / natural-language queries✓ Ask Your Data with pandas tool-useEvery answer is auditableQ&A and Copilot — improving but variable quality, costs extra One-click DOCX / PPTX export✓ Auto Report — editable Word + PowerPointPowerPoint export is decent; Word export is not native Forecasting✓ Holt-Winters with confidence bands✓ ETS forecasting (similar) RFM segmentation✓ Custom segments + templatesPossible via DAX measures (manual, complex) Anomaly detection✓ One-click on any time-series✓ Built-in anomaly detection Live database connectionsLimited (CSV / Excel / Sheets / Shopify)✓ Strong — 100+ connectors Data residency✓ UK / EU onlyConfigurable, default Microsoft cloud regions Setup time (CSV → first chart)~2 minutes~20–40 minutes (install Desktop, sign in, model, publish) ## Where DataHub Pro is genuinely better ### Pricing math without an Excel spreadsheet Power BI's pricing is famously hard to model. Pro at $14 is the headline, but that doesn't include row-level security, paginated reports, or advanced AI features (Premium per user at $24, or capacity at $5000+/mo). For a 5-person team that wants the "real" Power BI experience, you're often looking at $120+/user/mo. - 5 × Power BI Premium per user @ $24 = $120/month ≈ £95 - 5 × DataHub Pro Pro @ £19 = £95/month — same total, more clarity Add capacity-based costs and the gap widens fast. ### Cross-platform without compromise Power BI Desktop is the only tool that builds reports — and it's Windows-only. If your team is on Macs (most agencies, most modern startups), you're either dual-booting, running a VM, or stuck on the web experience that's missing key features like Power Query GUI, paginated reports, and complex DAX modelling. DataHub Pro runs in any browser on any device with the same feature set. Build dashboards from your iPad on a flight, edit on a Mac at your desk, share to a colleague's Linux laptop — all the same tool. ### AI that uses real data, not summarised data Power BI's natural-language Q&A and Copilot run against your data model. They're working with whatever measures and aggregations you've pre-built, which means questions outside that model often don't get answered. DataHub Pro's Ask Your Data runs real pandas operations directly on your file using a tool-use loop. Every answer comes with the tool calls that produced it (load_file_data → filter_by_date → aggregate_by_account) so the maths is auditable, not generated. ### Reports that go in a deck, not just a workspace Power BI's strength is the workspace experience — interactive dashboards your team browses. But most analyst output ends up in a Word document or a PowerPoint slide for a client or board. DataHub Pro's Auto Report generates a fully editable DOCX or PPTX in one click — title page, executive summary, charts, recommendations, all editable in Word/PowerPoint. Power BI exports to PDF and PowerPoint, but the PowerPoint export is image-based, not editable. ### When Power BI is still the right choice If most of these describe you, stick with Power BI: - You're a Microsoft 365 / Azure shop. Power BI's deep integration with Excel, SharePoint, Teams, Dynamics, Synapse, Azure SQL is genuinely best-in-class. The licence is often bundled. - You have a data team that knows DAX. DAX is powerful and the talent pool is huge. Don't throw that away. - You need row-level security with AD/Entra. Power BI's enterprise security model with Active Directory groups is mature and audited. - You need on-premises gateway connections to legacy databases inside corporate firewalls. - You're embedding analytics inside another Microsoft product. Power BI Embedded is the path of least resistance. If none of those describe your situation — and you're really just trying to ship a report off a CSV without booting Windows — DataHub Pro will save you money and time. ## Who DataHub Pro is built for The teams who get the most out of DataHub Pro instead of Power BI are typically: - Mac-first agencies and startups who don't want to dual-boot just to build a dashboard. - SMEs not on Microsoft 365 — Google Workspace shops, Notion-natives, anyone who'd be paying for a Microsoft license only for Power BI. - Consultants and freelancers serving multiple clients, where you don't want to manage Microsoft licenses across each engagement. - Founders and operators who'd rather ask "why did revenue spike on March 14?" than learn DAX. ## FAQs Is Power BI free really free? There's a free tier, but you can't share reports with anyone outside your tenant unless they also have Power BI accounts, and many enterprise features (RLS, paginated reports, large datasets) are gated behind Premium. DataHub Pro's free tier lets you upload files, build dashboards, and share with anyone via a public link. Can DataHub Pro run on Mac? Yes. DataHub Pro runs entirely in the browser. Power BI Desktop, the authoring tool, is Windows-only — Mac users have to use the much more limited web experience or run Windows in a VM. How does AI in DataHub Pro compare to Microsoft Copilot for Power BI? DataHub Pro's Ask Your Data runs real pandas operations on the underlying file with tool-use traced. Microsoft Copilot in Power BI runs against your pre-built data model — it's good at querying what you've already modeled, less good at exploratory questions outside it. Copilot also requires Premium licensing (~$24/user/mo extra). What about exports — does DataHub Pro do PowerPoint? Yes, both DOCX (Word) and PPTX (PowerPoint), as fully editable files in one click. Power BI exports to PowerPoint as static images embedded in slides, which means you can't edit the chart in PowerPoint after export. Can I migrate my Power BI reports? Not directly — DataHub Pro is built around CSV/Excel uploads rather than .pbix workbook files. Most teams switching from Power BI export the underlying data to CSV/Excel and rebuild dashboards in DataHub Pro. The dashboard builder takes minutes for typical reports. Does DataHub Pro work with Excel files? Yes. Drop in any .xlsx file — DataHub Pro reads it natively, infers types, suggests cleaning operations, and builds a dashboard automatically. What about data security and GDPR? DataHub Pro is UK/EU-first with all data on UK/EU infrastructure, never used for AI training. Full details in our Privacy Policy and GDPR & DPA page; signed DPA available on request. ## See it on your own data in 2 minutes. The free tier doesn't ask for a credit card. Drop in a CSV and the dashboard, insights, and report are ready before your coffee cools. Start free → Watch 2-min demo ✓ No credit card   ✓ 14-day full-access   ✓ Cancel any time Compare DataHub Pro to: Tableau · Domo · Looker · Looker Studio · Qlik Sense · Databox · Geckoboard · Mode · Metabase · Causal · Rows · Numerous AI · Sigma Further reading: Power BI on Wikipedia · Gartner BI reviews Compare with the rest of the market — we’ve published an independent 2026 review of all 12 AI tools for Excel data analysis (Ajelix, Formula Bot, GPT Excel, Microsoft Copilot, Numerous AI, ChatGPT, Polymer, Tableau, Power BI, Domo, KNIME). --- ## DataHub Pro vs Databox Source: https://www.datahubpro.co.uk/vs/databox Home/Compare/Databox Comparison · Updated May 2026 # The Databox alternative for teams who want flat-rate pricing and editable exports. Dr Waqas Rafique · Founder & CTO LinkedIn · About Databox is a polished live-dashboard tool that pulls from 100+ marketing APIs — fine if you want exactly that at $79+/month. DataHub Pro is broader: any CSV from any source, AI insights, RFM and forecasting, full white-label, and editable DOCX/PPTX exports. £19 per seat, flat — not per dashboard or per data source. ## Quick answer · What is the best Databox alternative for marketing agencies? DataHub Pro is a Databox alternative with flat per-seat pricing (£19/user/month) instead of per-dashboard or per-source. Reads any CSV from any platform, adds AI insights and Holt-Winters forecasting, and exports to editable Word/PowerPoint. Best for agencies serving 15+ clients where Databox's tier pricing scales aggressively. Start free trial → Watch 2-min demo ✓ No credit card   ✓ 14-day full-access   ✓ Cancel any time £19/seat Flat-rate per user vs Databox $79 starter / $399+ for premium Any CSV From any platform Not limited to pre-built API connectors DOCX/PPTX Editable exports Databox exports to PDF only ## Why teams switch from Databox to DataHub Pro If you're searching for a Databox alternative you usually fall into one of three buckets. Here's what we hear most often from teams who've made the move. 1 ### You report to clients in Word/PowerPoint, not screens Databox is great for live screens but produces PDF exports only. DataHub Pro generates fully editable Word documents and PowerPoint decks with your branding — what clients actually want for board meetings. 2 ### You have data outside Databox's connectors Databox covers most marketing APIs but not custom CRMs, hand-maintained Excel files, or warehouse exports. DataHub Pro reads any CSV — universal source compatibility. 3 ### Pricing punishes growth Databox per-dashboard / per-source pricing scales aggressively. DataHub Pro is flat £19/seat regardless of how many clients, dashboards or data sources you have. ## Side-by-side comparison Honest, feature-by-feature. Pricing accurate as of May 2026 based on each vendor's published rates (or, where pricing is custom, our best estimate from public sources and our own conversations). DataHub Pro Databox Starting price£19/user/mo$79/mo starter, $399+ premium Free tier✓ Yes, no credit cardFree tier exists but limited connectors Data sourcesAny CSV from any platform100+ pre-built connectors; CSV upload limited AI insights / narrative✓ Built in (Ask Your Data, AutoInsights)Limited AI commentary Editable Word/PowerPoint exports✓ One click DOCX + PPTX✗ PDF only Forecasting (Holt-Winters + bands)✓Limited forecast RFM customer segmentation✓✗ Anomaly detection✓ Auto on every numeric columnGoal-tracking but not statistical anomalies White-label per client✓ Per-client branding profiles✓ Available, additional cost Live API connectionsLimited — file/CSV-first✓ Strong Setup time~2 minutes10-30 minutes per source connector ## Where DataHub Pro is genuinely better ### Universal source compatibility Databox's strength is its 100+ pre-built connectors. The catch: if your data isn't in one of those connectors, you're stuck. DataHub Pro takes any CSV from any platform — Google Ads, Meta, GA4, Klaviyo, HubSpot, Salesforce, Stripe, Shopify, hand-maintained Excel, custom warehouse export. Universal source compatibility, no per-connector pricing. ### Editable Word and PowerPoint reports Databox exports to PDF — fine for archive, useless when your client wants to edit before sending. DataHub Pro generates fully editable DOCX and PPTX with your client's branding. Real Word tables, real PowerPoint slides, ready for the client to red-line. ### Deeper analyses out of the box Databox is dashboard-first. DataHub Pro adds: Holt-Winters forecasting with confidence bands, RFM customer segmentation, anomaly detection on every numeric column, cohort analysis, period comparison. The kind of analyses your clients actually want in a quarterly review, not just a live screen. ### Flat-rate pricing Databox starts at $79 and ramps fast as you add dashboards, sources, or users. DataHub Pro is flat £19 per agency seat. A 5-person agency serving 30 clients pays £95/month total on DataHub Pro vs $400+ on Databox premium tiers. ### When Databox is still the right choice If most of these describe you, Databox fits better: - Your data only lives in marketing platforms with native Databox connectors - You need real-time refresh (every minute) rather than weekly/monthly cadence - You want TV-style live dashboards for an office wall - Your clients want PDF reports, not editable docs For weekly/monthly client reporting cadence with editable deliverables, DataHub Pro is the better fit. ## Who DataHub Pro is built for Marketing agencies, performance marketing teams, in-house analytics functions at SMBs. The reporting layer for teams whose data lives in CSVs from many sources rather than one warehouse. ## FAQs How does the pricing actually compare for a 5-person agency with 30 clients? DataHub Pro: 5 seats × £19 = £95/month. Databox premium tier with that volume: $399-$799/month. Roughly 4-8× cheaper at this scale. Can I migrate my Databox dashboards to DataHub Pro? Not directly — different data models. The practical migration: export your underlying CSVs from each source, drop them into DataHub Pro, rebuild your KPI dashboard in 5-15 minutes. Most agencies report the rebuild is faster than the original Databox setup. Does DataHub Pro do live API pulls like Databox? Live API connectors are roadmap. Today: drop in CSV exports. Most agencies prefer this for audit reasons — snapshot files are reproducible, live links can't be. What about custom widgets / custom KPIs? Calculated columns and custom KPIs are first-class. Formula builder uses a real expression engine. White-label per-client report templates supported on Pro. Is white-label included or extra? White-label per-client is included on Pro tier — no upcharge. Databox tiers it (white-label is a Premium feature on their plans). How does AI compare? DataHub Pro's tool-use AI runs deterministic pandas operations on your data — every answer ships with the call trace. Databox's AI is generative-narrative summarisation. Different models; ours is auditable, theirs is faster but harder to verify. ## See it on your own data in 2 minutes. The free tier doesn't ask for a credit card. Drop in a CSV and the dashboard, insights, and report are ready before your coffee cools. Start free → Watch 2-min demo ✓ No credit card   ✓ 14-day full-access   ✓ Cancel any time Compare DataHub Pro to: Tableau · Power BI · Domo · Looker · Looker Studio · Qlik Sense · Geckoboard · Mode · Metabase · Causal · Rows · Numerous AI · Sigma Further reading: Business dashboards on Wikipedia · Databox reviews on G2 Compare with the rest of the market — we’ve published an independent 2026 review of all 12 AI tools for Excel data analysis (Ajelix, Formula Bot, GPT Excel, Microsoft Copilot, Numerous AI, ChatGPT, Polymer, Tableau, Power BI, Domo, KNIME). --- ## Holt-Winters forecasting overview Source: https://www.datahubpro.co.uk/holt-winters-forecasting Home/Features/Holt-Winters forecasting Feature · Forecasting # Holt-Winters forecasting. With confidence bands. Without the statistics PhD. Dr Waqas Rafique · Founder & CTO LinkedIn · About A Holt-Winters forecasting tool that doesn't make you read the original 1960 paper. DataHub Pro takes any time-series CSV — daily, weekly, monthly — fits a Holt-Winters model with trend and seasonality, and produces a forecast with 80% and 95% confidence bands. A what-if calculator lets you flex one driver (growth rate, seasonality strength, holdout periods) and see the impact across the forecast horizon. The whole flow runs without code, statistics packages, or Excel macros. ## Quick answer · What is Holt-Winters forecasting? Holt-Winters forecasting is an exponential-smoothing time-series method that decomposes data into level, trend and seasonal components and projects them forward, with auto-tuned smoothing parameters and confidence bands. It's the workhorse of business forecasting since 1960. DataHub Pro fits Holt-Winters automatically and produces 80%/95% confidence band charts in two minutes. Updated 7 May 2026 Start free trial → Watch 2-min demo ✓ No credit card   ✓ 14-day full-access   ✓ Cancel any time 80/95% Confidence bands on every forecast Visualises uncertainty so you don't trust point estimates Auto-tuned Trend, seasonality and damping Backtested against historical periods for accuracy What-if Sensitivity analysis with sliders Flex assumptions and see the impact instantly ## Why most forecasting tools are either too simple or too complex Time-series forecasting in business has two failure modes. ### Too simple · the LINEST trap Excel's straight-line LINEST is what most teams reach for. It assumes the future will look like a linear extension of the past — no trend curvature, no seasonality, no shocks. For most businesses (where Christmas exists, where summer is slower, where growth slows over time) it's wrong before it starts. ### Too complex · the Python statsmodels rabbit hole The other extreme is opening Python, fitting an ETS model with statsmodels, debugging frequency strings, choosing seasonal_periods, validating with pmdarima. This produces a great forecast but takes the kind of statistical confidence that most finance and ops teams don't have. ### Holt-Winters: the right middle ground Holt-Winters exponential smoothing handles trend + seasonal + residual decomposition with three intuitive parameters (level, trend, seasonal smoothing). It's been the workhorse of business forecasting for 60 years for good reason — it captures the patterns that matter without overfitting to noise. DataHub Pro tunes the parameters automatically and surfaces them so you can see what the model thinks. ## How Holt-Winters works in DataHub Pro Six layers around the core Holt-Winters fit. You don't need to use any of them — but they're there when you do. 1 ### Auto frequency detection The platform detects whether your data is daily, weekly, monthly or quarterly. Picks an appropriate seasonal period (7, 52, 12, 4) automatically. Override if your business has unusual seasonality. 2 ### Auto-tuned smoothing parameters Level (α), trend (β), and seasonal (γ) smoothing parameters are fit by minimising one-step-ahead error on a hold-out period. Tunable manually if you have domain knowledge. 3 ### Damped trends For long-horizon forecasts, the damped trend variant (Holt-Winters with φ damping) prevents trend extrapolation from running away. Auto-applied for forecasts >12 periods. 4 ### 80% and 95% confidence bands Computed from the residual variance and forecast horizon. Visualises uncertainty so you can see when the model is confident vs guessing. 5 ### Backtest mode Hold out the last N periods, fit on the rest, predict the held-out period, compute MAPE / MAE / RMSE. Tells you how the model would have done last quarter — useful before trusting it for next quarter. 6 ### What-if sensitivity slider Flex one driver (assumed growth rate, seasonality strength, etc.) and see the impact across the forecast. Quick way to bracket scenarios for board / lender meetings. ## What Holt-Winters forecasting unlocks - Cash-flow planning. Forecast 12 months ahead with confidence bands. Plan runway with the upper and lower bounds, not a single point. - Revenue forecasting. Capture seasonality (Q4 spikes, summer dips) automatically. Compare forecast vs actuals each month to refine. - Inventory planning. Daily-level forecast for stock-sensitive products. Confidence bands inform safety stock. - Headcount planning. Headcount required to deliver forecast revenue, with sensitivity to productivity assumptions. - Variance analysis. Forecast each line item, compare to actuals, flag the biggest variances. Each of these is a Holt-Winters fit + a confidence band + a what-if calculator — and that's most of what business forecasting actually needs. ## Who needs Holt-Winters forecasting ### Finance teams running cash-flow forecasts The classic use case. Drop in monthly cash flow; get 12-month forecast with confidence bands. Use upper/lower bounds for runway planning. ### Operations and ops research teams Inventory, capacity, throughput, demand. Daily or weekly granularity. Confidence bands inform reorder points and safety stock. ### Founders pitching investors Forecasts with confidence bands look more credible than single-line projections. Show the model's uncertainty rather than pretending it doesn't exist. ### When Holt-Winters is the wrong model Holt-Winters assumes the underlying process has stable seasonal patterns. If your business has structural breaks (a pivot, a major customer loss, a regulatory change), the model will keep extrapolating the old pattern. For these cases, segment the data into pre-break and post-break and forecast each separately, or use a more flexible model (Prophet, ARIMA with intervention). The platform supports manual segmentation; we're working on automatic break detection. ## FAQs What is Holt-Winters forecasting? Holt-Winters is a time-series forecasting method that decomposes a series into level, trend and seasonal components, smooths each with exponential weights, and projects them forward. It's been the workhorse of business forecasting since the 1960s — well-understood, robust, and accurate enough for most business decisions without requiring a statistician. Do I need to know statistics to use it? No. The platform fits the model automatically — picks parameters, validates on hold-out data, surfaces a forecast with confidence bands. You see the parameters if you want them, but the defaults are appropriate for most business time series. How accurate is it? For typical business time series with clear seasonality, Holt-Winters produces forecasts with MAPE in the 5-15% range — accurate enough for most planning decisions. Backtest mode shows you the actual MAPE on your data before you trust the forward forecast. Can it handle daily data? Yes. Daily, weekly, monthly, quarterly, yearly. Frequency is auto-detected; seasonality period is auto-set (7 for daily, 52 for weekly, 12 for monthly, 4 for quarterly). Override if your business has unusual seasonality patterns. What if my data has trend curvature, not just linear? Holt-Winters fits a damped trend by default for long-horizon forecasts, which handles trend curvature. For very strong non-linearity, you'd use a different model (we offer regression-based forecasting too). For most business data, damped Holt-Winters is the right choice. Does it work with multiple time series at once? Yes — multi-series forecasting. Drop in a CSV with multiple columns of time series; the platform fits each independently and produces a combined forecast view. Useful for product-level or region-level forecasts. How are confidence bands computed? From the residual variance of the fitted model and the forecast horizon. The further out you forecast, the wider the confidence band — reflecting that uncertainty grows with horizon. The 95% band means we expect the actual to fall inside the band 95% of the time, all else equal. Can I export the forecast to Excel? Yes — XLSX export with the forecast values, confidence band bounds, and any what-if scenarios you ran. Editable in Excel for downstream modelling. Is the what-if calculator easy to use? Yes — slider-based UI. Flex assumed growth rate, seasonality strength, or apply a one-time shock and see the forecast update instantly. Useful for bracketing scenarios for board meetings. How does this compare to Prophet (Facebook's forecasting tool)? Prophet is excellent for time series with strong holiday effects and changepoints. Holt-Winters is more robust for typical business data without exotic effects, faster to fit, and easier to interpret. We've found Holt-Winters wins for most finance and ops use cases; Prophet wins for retail with strong holidays. Both are good tools; Holt-Winters is the right default. ## Forecast your next 12 months with confidence bands. Drop in your historic monthly numbers. Get a Holt-Winters forecast with 80/95% bands in seconds. Free tier — no credit card. Start free → Watch 2-min demo ✓ No credit card   ✓ 14-day full-access   ✓ Cancel any time See also: · For accountants · For finance teams · For insight agencies · For small business · · Home Further reading: Exponential smoothing on Wikipedia · Hyndman & Athanasopoulos: Forecasting Principles and Practice Compare with the rest of the market — we’ve published an independent 2026 review of all 12 AI tools for Excel data analysis (Ajelix, Formula Bot, GPT Excel, Microsoft Copilot, Numerous AI, ChatGPT, Polymer, Tableau, Power BI, Domo, KNIME). --- ## RFM segmentation overview Source: https://www.datahubpro.co.uk/rfm-segmentation Home/Features/RFM segmentation Feature · Customer segmentation # RFM segmentation in two minutes, not two days. Dr Waqas Rafique · Founder & CTO LinkedIn · About An RFM segmentation tool done properly: drop in any transaction CSV — customer, date, amount — and get Recency / Frequency / Monetary scores, four tunable segments (Champions, Loyal, At-Risk, Lost), and a per-segment action list with retention recommendations and CSV export. No SQL, no spreadsheet macro, no agency hand-off. ## Quick answer · What is RFM segmentation? RFM segmentation is a customer scoring method that ranks each customer 1-5 on Recency, Frequency and Monetary value, then maps the combined score to named tiers (Champions, Loyal, At-Risk, Lost). DataHub Pro runs RFM on any transaction CSV in two minutes — segments, retention actions, and CRM-ready CSV export included. Updated 7 May 2026 Start free trial → Watch 2-min demo ✓ No credit card   ✓ 14-day full-access   ✓ Cancel any time 2 min Transaction CSV → segments → CSV export vs days of manual SQL or pivot-table work Tunable Segment thresholds + custom segments Default 4 segments fit most businesses; override anything AI actions Per-segment retention recommendations Backed by real customer behaviour, not generic templates ## Why most teams skip RFM analysis RFM (Recency / Frequency / Monetary) is one of the most powerful customer segmentations in business — and one of the least-used. The reason is friction. ### The traditional RFM workflow - Pull transaction history from your data warehouse (or join CSVs from billing + CRM). - Compute per-customer Recency, Frequency, Monetary metrics. - Bucket each into quintiles (1-5). - Combine into RFM scores; map to named segments. - Build a dashboard. Build an action list. Get the marketing team's input. Steps 1 and 2 alone take half a day for most analysts. By the time the segments are ready, the marketing campaign window has closed. ### What RFM segmentation should be Drop in transaction history. Get segments. Get actions. Take the actions before the customer churns. ## How RFM segmentation works in DataHub Pro Six layers — each one shaves time off the traditional workflow. 1 ### Auto column detection The platform identifies your customer ID, transaction date and amount columns automatically. Override if the auto-detect picks the wrong columns. 2 ### Quintile scoring Recency: days since last transaction → 1–5 score (5 = most recent). Frequency: number of transactions → 1–5. Monetary: total spend → 1–5. Combined into a 3-digit RFM score per customer. 3 ### Default 4-segment mapping Champions (high R, F, M), Loyal (medium R, high F+M), At-Risk (low R, high prior F+M), Lost (low R, F, M). Sensible defaults that match how most businesses think about their base. 4 ### Custom segments Define your own segments by score combination. "New & promising" (high R, low F), "Whales" (top 5% Monetary), "Discount-only" (high F, low M). Combine with metadata for advanced cuts. 5 ### AI-generated retention actions For each segment, the AI proposes 3-5 retention actions based on the segment's behaviour and your industry. Editable. Linked to typical channels (email, SMS, sales rep, ads). 6 ### CSV export and CRM sync Export each segment as a CSV (customer ID, contact info if present, RFM score, action recommended). Drop into your CRM, email tool, or ad platform for targeting. CRM-direct sync on the roadmap. ## What teams do with RFM segmentation - At-Risk re-engagement. Identify customers who used to spend regularly but haven't recently. Trigger a reactivation campaign before they churn entirely. - Champions retention. Top 5% by Monetary often drive 30-50% of revenue. Treat them as their own segment — early access, account management, NPS check-ins. - Loyal upsell. High frequency, medium spend. Test premium tier or expansion offers — this segment has the highest conversion to upsell. - New cohort onboarding. High recency, low frequency. Onboarding campaign to drive second purchase before the "new customer" window closes. - Lost recovery. Low across the board. Selective win-back campaign with a strong offer; if no response, move to suppression list to lower acquisition CAC inflation. Each of these is a 5-10 minute job once the RFM segments are built. ## Who RFM segmentation suits ### E-commerce founders and marketers The classic fit. Transaction data is structured (customer, order, date, total), which is exactly what RFM needs. Most e-commerce teams who run RFM properly see a 10-30% lift in 90-day retention. ### SaaS retention teams Recency = last login or last meaningful action. Frequency = active days per month. Monetary = MRR. Same model, slightly different definitions; same insight into who needs what. ### B2B sales teams Customer = account; Recency = last order; Frequency = orders per quarter; Monetary = ARR or rolling 12-month revenue. Top accounts get account management; at-risk accounts get a CSM call. ### Where RFM hits its limits RFM is descriptive, not predictive. It tells you who looks at-risk now based on past behaviour, but it doesn't predict who will churn next month based on signals like product usage, support tickets, or sentiment. For predictive churn modelling, layer the platform's Churn Risk feature on top — it uses RFM as one of several signals. ## FAQs What does RFM stand for? Recency, Frequency, Monetary. Three dimensions of customer value: how recently they bought, how often they buy, and how much they spend. The combination of all three predicts retention and lifetime value better than any of them alone. How is RFM scored in DataHub Pro? Quintile scoring on each dimension: customers are bucketed into 1-5 (5 = best). The combined RFM score is a 3-digit code (e.g. 555 = Champion, 111 = Lost). The four default segments map score ranges to human-readable tiers. Can I use it for B2B with longer purchase cycles? Yes. The default time windows fit B2C / e-commerce; for B2B you'd typically widen the recency window (e.g. 90 days vs 30) and use ARR or rolling 12-month revenue for Monetary. The platform exposes these parameters; defaults work for most businesses without changes. How is "at-risk" defined? By default, customers who scored high on Frequency and Monetary in the past but have low Recency now — meaning they used to be valuable but haven't bought lately. The exact threshold is tunable; most teams find the default surfaces the right cohort. Can I integrate with my CRM or email tool? CSV export for any segment is one click. Direct sync to HubSpot, Salesforce, Mailchimp, Klaviyo is on the roadmap. Most teams export to CSV and import into their tool of choice. Does it support custom segments beyond the default four? Yes. Custom Segments lets you define any combination of R, F, M score ranges. "New & promising", "Discount-only", "Whales" — name your own segments based on the cuts that matter to your business. How big a transaction file can I run RFM on? Free tier: 100,000 transactions. Pro tier: 2 million. Beyond that, you'd pre-aggregate to per-customer summaries before uploading. Does the AI generate the retention recommendations? Yes. For each segment the platform proposes 3-5 actions based on the segment's behaviour pattern (and your industry context if known). The actions are editable; you'd typically refine them to match your tone of voice and existing campaigns. Can I track segment transitions over time? Yes — re-run RFM month over month and the platform tracks segment movement (Loyal → Champion, At-Risk → Lost). Useful for measuring whether retention campaigns actually moved the customer back up the value tiers. How does RFM relate to LTV and CAC? RFM segments correlate strongly with future LTV (Champions and Loyal tiers usually generate 5-10× the LTV of Lost and At-Risk). Combined with CAC, RFM tells you which acquisition channels and campaigns produce customers in the higher-value segments — actionable input to spend optimisation. ## Run your first RFM segmentation in 2 minutes. Drop in your transaction CSV. Get four customer tiers and an action list. Free tier — no credit card. Start free → Watch 2-min demo ✓ No credit card   ✓ 14-day full-access   ✓ Cancel any time See also: · For accountants · For finance teams · For insight agencies · For small business · · Home Further reading: RFM on Wikipedia · Customer lifetime value on Wikipedia Compare with the rest of the market — we’ve published an independent 2026 review of all 12 AI tools for Excel data analysis (Ajelix, Formula Bot, GPT Excel, Microsoft Copilot, Numerous AI, ChatGPT, Polymer, Tableau, Power BI, Domo, KNIME). --- ## Anomaly detection in Excel Source: https://www.datahubpro.co.uk/anomaly-detection-excel Home/Features/Anomaly detection for Excel Feature · Anomaly detection # Anomaly detection for Excel. Outliers caught before the board does. Dr Waqas Rafique · Founder & CTO LinkedIn · About Anomaly detection for Excel usually means "eyeball the chart and hope". DataHub Pro scans every numeric column on upload using rolling z-scores and surfaces points that are statistically out of distribution — with the date, the expected range, and how many standard deviations they sit at. Useful for variance reviews, audit prep, reconciliation work, and catching the data-quality issue before it lands in the board pack. ## Quick answer · How do I detect anomalies in Excel? Detect anomalies in Excel by uploading the file to DataHub Pro — every numeric column is scanned with rolling z-scores, with points beyond 2 or 3 standard deviations flagged as warnings or critical. Each anomaly comes with the date, value, expected range and click-through to source rows, ready for variance review or audit prep. Updated 7 May 2026 Start free trial → Watch 2-min demo ✓ No credit card   ✓ 14-day full-access   ✓ Cancel any time Every column All numeric columns scanned Revenue, costs, headcount, orders — anything numeric Z-scores Rolling 12-period mean + stdev Configurable window; robust to slowly-changing data Click-through Each anomaly to source rows Drill from the flag to the underlying data ## Why anomaly detection isn't standard in Excel reporting Most teams know they should be checking for outliers in their data — strange months, missed accruals, coding errors, fraud. Most teams don't actually do it because the workflow is painful. ### The traditional approach - Build a chart of the time series. - Eyeball it for points that "look weird". - Build a simple z-score column with =STDEV / =AVERAGE. - Compute thresholds manually. Apply conditional formatting. - Repeat for every column you care about. For one column, doable. For a 30-column trial balance, untenable. So most teams skip it and discover the anomaly when someone notices in a meeting two weeks later. ### What anomaly detection should be Drop in the spreadsheet. Every numeric column is scanned. Anomalies are flagged with date, value, expected range, severity. Click any flag to see the source rows. Fold into your variance review or audit pack. ## How DataHub Pro flags anomalies Six layers around the core rolling-z-score detection. Each adds context that makes the alerts actionable. 1 ### Rolling z-score on every column Default 12-period rolling mean + stdev. Each new point's z-score (deviations from rolling mean) is computed; |z| > 2 flagged as warning, |z| > 3 as critical. Tunable thresholds and window. 2 ### Auto-applied to all numeric columns On upload the platform scans every numeric column for anomalies. No need to specify which to check; the dashboard surfaces the top flags across the whole file. 3 ### Severity ranking Anomalies ranked by absolute z-score and value impact. The top 5 are surfaced first; full list available on click. Useful for prioritising which to investigate. 4 ### Missing-period detection Beyond outlier values, the platform detects missing time periods (e.g. a quarter with no data). Useful for catching coding gaps or system-export failures. 5 ### Click-through to source rows Each anomaly links to the underlying rows. See the transactions that drove the spike. Speeds up investigation from minutes to seconds. 6 ### Inclusion in Auto Report Anomalies surface in the Auto Report DOCX/PPTX as a dedicated section with severity, value and explanation. Stops material outliers being missed in the executive summary. ## Where anomaly detection earns its keep - Variance review. The biggest variances vs budget are usually anomalies. The platform flags them in the same view, with the source rows one click away. - Audit prep / Y/E. Top anomalies become the audit work programme. Document the explanation for each in the working papers. - Reconciliation. Differences between systems often show as anomalies in one of the two. Catch them in the file load rather than the reconciliation step. - Fraud detection. Statistical outliers (transactions far from the rolling average) are a signal — not proof, but a useful filter for human review. - Data quality. Missing-period detection catches export failures, system gaps, missed cutoffs. Better caught at upload than at month-end. None of these are exotic uses — they're the everyday work that takes hours of manual chart-staring and minutes of rolling z-score. ## Who needs anomaly detection in spreadsheets ### Finance, FP&A, controllers Variance analysis, month-end review, audit prep. Anomalies in the trial balance catch coding errors before they reach the audit; anomalies in cost lines catch missed accruals; anomalies in revenue catch unusual one-off items. ### Internal audit and compliance Statistical sampling for review. Anomalies in transaction data surface candidates for deeper investigation. Documented in working papers with the rationale. ### Operations and revenue analysts KPIs that are out of distribution often point to a problem (or an opportunity). Daily/weekly KPI reviews with anomaly flags surface what to investigate. ### What anomaly detection won't catch Statistical outliers aren't the same as "wrong" numbers. A genuinely large month (a one-off contract, a windfall sale) flags as an anomaly even though the data is correct. The platform's job is to surface anomalies for human review, not to decide what's wrong. Expect ~5-10% of high-z-score flags to be legitimate large months that needed surfacing for the commentary. ## FAQs What threshold defines an anomaly? Default: |z-score| > 2 for warning (≈5% of points in normal distribution), |z-score| > 3 for critical (≈0.3% of points). Tunable in the settings; some teams use ±2.5 to balance recall and noise. Can I configure the rolling window? Yes. Default 12 periods (one year of monthly data; one quarter of weekly data). Tune to match your business cycle. Shorter windows are more responsive to recent changes; longer windows are more stable. Does it work on financial data with strong seasonality? Yes — but for highly seasonal data (think retail with a 3× Q4 spike), a deseasonalised z-score is more useful than the raw rolling z-score. The platform applies seasonal decomposition where it detects clear seasonality and computes z-scores on the residuals. How does it differ from Excel's conditional formatting? Conditional formatting highlights cells based on absolute thresholds you specify. Anomaly detection computes statistical thresholds based on the rolling distribution, so a number that's anomalous now might not be anomalous in three months as the underlying mean shifts. The dynamic threshold means false positives are lower and the alerts are more useful. Can I exclude known one-offs? Yes. Mark a flagged anomaly as "explained" with a note (e.g. "one-off Christmas bonus pool") and it stops being surfaced. The audit log retains the original flag and the explanation. Does it work on revenue, costs, headcount, etc.? Anything numeric. The platform doesn't make domain assumptions — z-scores are dimensionless. Revenue spikes, cost dips, headcount jumps are all flagged the same way. Will it produce too many alerts? The default thresholds are calibrated to surface roughly 1-3 anomalies per 100 points — typically the right volume for a human reviewer to assess in a few minutes. Tunable if your file has more or less "normal noise". Can it detect anomalies across multiple columns at once? Yes — "multivariate anomalies" (e.g. revenue and customer count moving in opposite directions) are detected via correlation residuals. Default mode is per-column z-score; advanced mode opens the multivariate view. How does anomaly detection appear in the Auto Report? A dedicated section: Anomalies Flagged. Lists severity, date, value, expected range, and a link to source rows. Editable text per anomaly so you can add the explanation before sending the report. Can I export anomalies as a CSV? Yes — every anomaly with its date, column, value, z-score and severity. Useful for working papers or pasting into a tracking spreadsheet. ## Catch the next outlier before the board does. Drop in your most recent finance or ops file. Every numeric column scanned. Free tier — no credit card. Start free → Watch 2-min demo ✓ No credit card   ✓ 14-day full-access   ✓ Cancel any time See also: · For accountants · For finance teams · For insight agencies · For small business · · Home Further reading: Anomaly detection on Wikipedia · Z-score on Wikipedia Compare with the rest of the market — we’ve published an independent 2026 review of all 12 AI tools for Excel data analysis (Ajelix, Formula Bot, GPT Excel, Microsoft Copilot, Numerous AI, ChatGPT, Polymer, Tableau, Power BI, Domo, KNIME). --- ## AI data analyst Source: https://www.datahubpro.co.uk/ai-data-analyst Home/Features/AI data analyst Feature · AI Analyst # An AI data analyst on tap — without the hallucinations. Dr Waqas Rafique · Founder & CTO LinkedIn · About An AI data analyst that doesn't fabricate numbers is harder to build than it sounds. DataHub Pro is one of a handful of products that gets this right: every analysis is produced by deterministic tools running on your file, every chart is real data, every report is editable Word or PowerPoint. The result is an analyst-shaped layer that handles the repetitive 80% of analytics — the upload, the cleaning, the KPI scan, the chart generation, the summary writing — so your humans can spend time on the 20% that actually needs judgement. ## Quick answer · What is an AI data analyst? An AI data analyst is software that handles the repetitive 80% of analytics — file ingest, cleaning, KPI selection, chart generation, narrative summarisation, deliverable production — so humans focus on judgment work. DataHub Pro covers the full job end-to-end on spreadsheet inputs, with editable Word/PowerPoint output for £19/user/month. Updated 7 May 2026 Start free trial → Watch 2-min demo ✓ No credit card   ✓ 14-day full-access   ✓ Cancel any time <2 min From CSV upload to a finished deliverable Insights, charts, KPIs, exportable Word/PowerPoint Auditable Every number ships with its tool trace Reproducible from the same input, every time £19/mo Per-user Pro tier Less than 1 hour of an analyst's time per month ## What an AI data analyst should actually do The phrase "AI data analyst" is overloaded — it's been applied to everything from in-cell Excel macros to multi-million-dollar enterprise platforms. To pick the right one, focus on what a junior analyst on a real team gets handed. ### The job, end to end - Receive a file. Usually a CSV from a system export, sometimes Excel from a colleague. - Clean it. Headers, types, missing values, duplicates, weird date formats. - Pick the analyses. KPIs, period comparison, segmentation, forecast — depending on what the file is. - Run the analyses. In Excel, in Python, or in the BI tool. - Write the summary. What changed, what's interesting, what to do next. - Build the deliverable. Word doc for clients, PowerPoint for the board, dashboard for the team. - Field follow-up questions. "Why did X drop?", "Can you cut this by region?" Most "AI analyst" tools cover steps 4 and 5 — they generate insights and prose. DataHub Pro covers steps 1 through 7, end to end, on the file you uploaded. ## How DataHub Pro covers the analyst job Six layers, each backed by code. The pipeline runs on upload; you intervene only where you want to. 1 ### File ingest + cleaning Multi-format ingest (.csv, .xlsx, .xls, .tsv), header detection, type inference, date parsing, duplicate detection, AI Cleanse for stuck files. Most files are usable within seconds; the few that aren't surface a one-prompt fix. 2 ### Analysis selection On upload, the platform identifies the file's shape (revenue + date = forecast candidate; customer + transaction = RFM candidate; actuals + budget = variance candidate) and suggests the right tools. 3 ### Analysis execution KPIs computed via deterministic aggregation. Forecasting via Holt-Winters with confidence intervals. Anomaly detection via rolling z-score. RFM scoring with tunable bins. Cohort retention with monthly cohorts. All real maths, no LLM in the maths path. 4 ### Narrative summarisation AutoInsights writes 3 plain-English insights from the computed numbers. Each insight cites the period and points used. Auto Report extends this to a full executive summary with recommendations. 5 ### Deliverable build Auto Report builds a fully editable DOCX and PPTX. Title page, exec summary, KPI tables, chart images, anomalies, recommendations, appendix. White-labelled with logos on Pro. 6 ### Follow-up questioning Ask Your Data lets stakeholders ask follow-up questions on the same file in natural language. Each answer comes with a tool trace so review is fast. ## What you offload to the AI analyst Map a typical analyst's week onto what the platform handles directly: - Monday — variance pack: upload actuals.xlsx + budget.xlsx → AI runs Period Comparison + Anomaly Detection → exports DOCX with executive summary. Was 4 hours, now 15 minutes. - Tuesday — pipeline review: upload Salesforce export → AI runs KPI dashboard, conversion rates, deal-aging segmentation → pinned dashboard for sales leadership. Was 2 hours, now 10 minutes. - Wednesday — client KPI report: upload client metrics → AI runs AutoInsights + AI Narrative + Auto Report (white-labelled) → DOCX ready to email. Was 3 hours, now 20 minutes. - Thursday — board pack: upload finance + ops + sales → AI runs unified KPIs and forecasts → PPTX with one section per function. Was a day, now an hour. - Friday — ad-hoc questions: stakeholders ask their own questions in Ask Your Data, with audit trails. Was 5+ interruptions, now self-serve. The human judgement — what to flag for the CFO, when to push back on the budget, what story to tell — doesn't go anywhere. The mechanical work does. ## Who an AI data analyst suits ### Single-analyst teams or no-analyst teams SMEs, growth-stage startups, agencies. If you have one analyst (or none) and a backlog of analysis requests, the platform is the cheapest way to expand capacity. ### Existing analytics teams who want to focus upstream Mature teams use the platform to offload routine reporting so they can focus on data strategy, modelling, instrumentation. The AI analyst handles tier-1 requests; humans handle tier-2 and tier-3. ### Insight agencies and consultancies White-label the AI's output with the client's logo. Ship the same volume of deliverables with smaller teams. Senior consultants spend their time on strategy rather than chart formatting. ### What the AI analyst can't do (yet) It can't sit in your data council and judge tradeoffs. It can't decide what your strategy should be. It can't notice the absence of data — the missing column you forgot to export. It can't replace the relationship between an analyst and a stakeholder. It does the mechanical 80% of analytics with high accuracy and a complete audit trail; the remaining 20% is still human work. ## FAQs How accurate is an AI data analyst on real client work? Numerical accuracy is essentially 100% on the deterministic operations — the maths is real Python on real data. Where errors occur, they tend to be in interpretation: AI picks the wrong column for an aggregation, or misreads a stakeholder's question. The audit trail surfaces these immediately, so a human review of an AI deliverable is much faster than building it from scratch. Can it replace a junior analyst? It can replace the mechanical part of a junior analyst's role — file cleaning, KPI computation, chart making, draft writing, repeat reporting. It doesn't replace the relationship-building, stakeholder management, or judgement work. Most teams find they can support 2-3× the analytics volume with the same headcount, rather than reducing headcount. Does it learn from my data over time? It learns workspace-specific conventions (your column names, your KPI definitions, your branding) within your account. It does not pool data across customers or train shared models on your inputs. UK/EU data residency, no model training on customer data, is the default. How does it compare to enterprise AI analyst tools (ThoughtSpot, Tableau Pulse)? Those products are warehouse-native — they sit on top of a Snowflake / BigQuery / Redshift connection and answer questions over live data. Strong fit for organisations with a real data warehouse. DataHub Pro is spreadsheet-native — its strength is the workflow that starts with a CSV/Excel file and ends with a Word doc or PowerPoint deck. Different shape, different price point. Can I integrate it with our other tools? Yes — Shopify, Google Sheets, SharePoint and FTP/S3 connectors are live. Stripe and HubSpot via CSV export. Webhooks and a public API are on the roadmap. For most workflows, a connector or a CSV export is enough. Will the output match our brand? Yes — Pro tier white-labelling lets you upload a logo, set a primary brand colour, and tune typography. DOCX and PPTX exports apply your branding throughout. Custom report templates (full design control) are an Enterprise feature. How does the AI analyst handle messy data? Three layers: (1) the parser handles common Excel/CSV oddities automatically (header detection, type coercion, date parsing); (2) the AI Cleanse tool takes natural-language cleaning instructions; (3) Pipelines (visual ETL) handle joins, unpivots, calculated fields. Most files are usable within minutes even if they arrive messy. Is there an audit trail for compliance? Yes — every AI tool call, every export, every share is logged. Audit logs are available to Owners and exportable as CSV. For regulated industries, the deterministic-tool design means each AI answer is reproducible from its input. How is this different from Excel's Copilot or Google's Gemini in Sheets? Both are excellent for in-sheet productivity (formula generation, summarising, formatting). DataHub Pro is built for the next step: turning the spreadsheet into a dashboard, an insight summary, and a Word/PowerPoint report. Many teams use both — Copilot inside Excel, DataHub Pro for the workflow that ends in a deliverable. What's the pricing? Free tier with limited access (no credit card). Pro at £19/user/month for unlimited analyses, white-label, scheduled reports, audit logs, larger files. Enterprise is custom for SSO, custom contracts, dedicated regions and SLA. See pricing. ## Hire an AI data analyst for £19/month. Free tier first — no credit card. Drop in a real file from this week's work and judge for yourself. Start free → Watch 2-min demo ✓ No credit card   ✓ 14-day full-access   ✓ Cancel any time See also: · AI spreadsheet analysis · AI Excel analysis · Ask Your Data · Automated reports from Excel · · Home Further reading: Data analysis on Wikipedia · Analyst role definition (BLS) Compare with the rest of the market — we’ve published an independent 2026 review of all 12 AI tools for Excel data analysis (Ajelix, Formula Bot, GPT Excel, Microsoft Copilot, Numerous AI, ChatGPT, Polymer, Tableau, Power BI, Domo, KNIME). --- ## Ask your data Source: https://www.datahubpro.co.uk/ask-your-data Home/Features/Ask Your Data Feature · Conversational Analytics # Ask Your Data. Get answers backed by real maths. Dr Waqas Rafique · Founder & CTO LinkedIn · About Ask Your Data in DataHub Pro is a tool-use AI: it picks the right pandas operation, runs it on your file, and returns the answer with the call trace attached. No more "the AI said the total was £4.2m, but it didn't actually compute the total". Every answer ships with the rows used, the filters applied, and the function called — so reviewing the AI's work is faster than redoing it. ## Quick answer · What is Ask Your Data? Ask Your Data is a tool-use AI that runs real pandas operations on your spreadsheet rather than predicting answers from text. Every numerical answer ships with the call trace — load_file_data, filter_by_date, aggregate_by_field — so the maths is reproducible and auditable. No hallucinated numbers, no truncated context. Updated 7 May 2026 Start free trial → Watch 2-min demo ✓ No credit card   ✓ 14-day full-access   ✓ Cancel any time 20+ Tools the AI can call load_file_data, filter_by_value, aggregate_by_field, correlate, etc. 100% Of answers come with a call trace Audit any number in seconds, not minutes 5–8s Median answer latency On files under 1M rows ## Why most "chat with your data" products miss The promise of a conversational analyst is genuinely compelling: type a question, get an answer. The reality with most current implementations falls into one of three traps. ### Trap 1 · Free-form text generation The model reads your data as text and predicts plausible answers. Sometimes those answers are right. Sometimes they're confidently wrong. You can't tell which without recomputing — defeating the point of the AI. ### Trap 2 · Pre-built dashboards in disguise Some tools route every "question" to a pre-built tile. Ask anything outside the tile and you get "I can't answer that". It's a glorified search bar, not a chat with your data. ### Trap 3 · No audit trail Even when the AI does compute, you don't see the working. For finance, audit, regulated industries — "trust me" isn't a deliverable. Ask Your Data is built on the opposite premise: the AI exists only to orchestrate tools. Every number it returns was produced by code on your data, and the code is shown. ## How tool-use AI works under the hood A loop, not a single LLM call. The model proposes a tool to call; the tool runs against your file; the result feeds back; the loop continues until the answer is grounded. 1 ### Tool catalogue, not free generation The AI can only act through 20+ deterministic tools — load file, head, describe, filter by value/date/regex, aggregate by group, sort, correlate, top-N, pivot, forecast, segment. Each tool returns numbers, not prose. 2 ### Visible reasoning trace Each AI turn shows: thought (one sentence on what to do next), tool name, arguments, result. Click any answer to expand the trace. The same question gives the same trace and the same answer every time. 3 ### Multi-step workflows Questions like "top 5 customers by revenue in Q3 minus refunds" require multiple tool calls — load → filter Q3 → aggregate by customer → subtract refunds → top 5. The loop handles the orchestration. 4 ### Conversation memory Follow-ups work naturally: "why did customer X drop in October?", "break that down by product", "forecast for the next quarter". The model carries context across turns and re-uses prior tool outputs where it can. 5 ### Charts on demand Ask for a chart in natural language and the AI picks an appropriate type (bar, line, scatter, histogram) and renders it inline. Charts can be pinned to a custom dashboard or sent into Auto Report. 6 ### Suggested follow-ups After every answer, the platform proposes 2-3 high-leverage follow-up questions based on what's interesting in your data. They're not generic prompts — they're contextual to the file you uploaded. ## What this looks like in practice Three example sessions, each starting from a single CSV upload. ### Session 1 · Sales analysis - You: "What's our biggest customer this year?" - Trace: filter_by_date(2026-01-01, today) → aggregate_by_field(customer, sum, revenue) → sort(desc) → top_n(1) - Answer: "Acme Ltd at £487,200 across 18 invoices." — with the underlying rows shown. - You: "Plot their monthly revenue." - → chart rendered inline, pinnable. ### Session 2 · Variance investigation - You: "Why was October so far over budget?" - Trace: filter_by_date(2025-10) → group_by(line_item) → variance(actual, budget) → sort(desc, abs) - Answer: "Two line items drove 86% of the variance: Marketing £42,100 over (campaign launch), and Cost of Sales £28,400 under (delayed supplier invoices)." ### Session 3 · Customer cohort - You: "Which clients haven't ordered for 90 days?" - Trace: aggregate_by_field(customer, max, order_date) → filter_by_date(< today - 90) - Answer: 47 clients, total prior LTV £312,000. CSV download offered. Each of these would have taken 5–20 minutes in Excel. Each took under 30 seconds in Ask Your Data, with a complete audit trail of what was computed. ## Who Ask Your Data is built for ### Analysts who get interrupted with ad-hoc questions The classic "just one quick number" from a stakeholder that takes 20 minutes to compute. Hand them a Ask Your Data session on the underlying file and they answer their own questions. ### Finance, audit and regulated teams The audit trail is the unlock. Any answer returned is reproducible from the trace, so internal review and external audit both move faster. ### Founders and operators without an analyst If you don't have a data team, Ask Your Data is the closest thing to having one on demand. Drop in your Stripe export or your CRM CSV and start asking questions. ### When Ask Your Data is the wrong tool If you need true real-time analytics over a transactional database (sub-second latency, freshness measured in seconds), this isn't that. Ask Your Data works on the file you upload at upload time. For live data, use a warehouse-connected BI tool — and bring the snapshot here when you need to write a report. ## FAQs How is this different from ChatGPT's Advanced Data Analysis? ChatGPT's ADA mode runs Python in a sandboxed container and can do real maths. The differences: (1) ADA generates code in the open ended chat — there's no curated tool catalogue, so a wrong column name silently produces a wrong answer; (2) the audit trail is the chat history, not a structured trace; (3) session timeouts and lost context are common; (4) there's no native export to Word/PowerPoint, white-labelling, or dashboard pinning. Ask Your Data is purpose-built for analytics workflows that end in a deliverable. Does the AI ever hallucinate numbers? Numerical answers come from deterministic tools running real pandas on your file. The model's role is to pick the tool, not to invent the number. We've not seen a confirmed hallucinated number in production, but the audit trail makes any error visible — wrong column picked, wrong filter applied — so you'd see it before acting on it. Can it answer questions across multiple sheets or files? Within a single workbook (multiple sheets), yes — pick a primary sheet on upload and the AI can join across sheets when asked. Across multiple files, the workflow is to combine them first using Pipelines (the platform's join/transform builder) and then chat over the combined output. We're working on multi-file native support. Does it remember previous sessions? Within a single conversation, yes — context, prior answers, and tool outputs are reused for follow-ups. Across sessions, no — each new conversation starts clean for data-isolation reasons. Saved insights and pinned charts persist across sessions in the dashboard. How long do answers take? Most answers return in 5–8 seconds on files under 1M rows. Multi-step questions (e.g. "top 5 customers by revenue minus refunds, plotted monthly") can take 15–25 seconds because each tool call is a round-trip. Latency on very large files (10M+) isn't supported on the spreadsheet pathway — those workflows belong on a warehouse. Can I see what data the AI looked at? Yes. Each tool call shows the rows it touched and the columns it used. For aggregations, you can expand the trace to see the underlying detail. For filters, the output row count is visible — so "the AI summed 50 rows" is never confused with "the AI summed 5,000 rows". Is there a free tier? Yes. The free tier includes Ask Your Data with a daily question limit. Pro removes the limit and adds export, dashboard pinning, scheduled questions and audit logs. See pricing for details. Can I use Ask Your Data on Shopify, Stripe, or HubSpot exports? Yes — any of those tools' CSV exports work directly. We also have native Shopify and Google Sheets connectors that pull data on a schedule, so the file in Ask Your Data is always fresh. Does the chat support charts? Yes. Ask "plot revenue by month" and the AI picks an appropriate chart type, renders it inline, and offers options to pin it to a dashboard or include it in Auto Report. Chart type can be overridden if you want a specific format. How is data handled in regulated industries? UK/EU data residency by default. SOC 2 Type II in progress. Signed DPA available on request. Audit logs of every AI conversation, tool call and export. We're a fit for finance, professional services, healthcare back-office and regulated SMEs; we're not yet certified for HIPAA-PHI workloads. ## Have a real conversation with your spreadsheet. Drop in a CSV. Ask the questions you'd usually ask a junior analyst. Get answers with audit trails attached, in seconds. Start free → Watch 2-min demo ✓ No credit card   ✓ 14-day full-access   ✓ Cancel any time See also: · AI spreadsheet analysis · AI Excel analysis · AI data analyst · Automated reports from Excel · · Home Further reading: OpenAI function-calling docs · AI hallucination on Wikipedia Compare with the rest of the market — we’ve published an independent 2026 review of all 12 AI tools for Excel data analysis (Ajelix, Formula Bot, GPT Excel, Microsoft Copilot, Numerous AI, ChatGPT, Polymer, Tableau, Power BI, Domo, KNIME). --- ## Changelog Source: https://www.datahubpro.co.uk/changelog # Changelog What we’re shipping. New tools and improvements published as they go live. Follow along by checking back — or grab the founder’s LinkedIn where releases get cross-posted. 9 May 2026 New ## Free in-browser cohort retention tool, founder press kit, and full credentialed E-E-A-T pass - New tool: Free cohort retention calculator — paste two columns of customer + date data, get the heatmap. No SQL, no signup, runs locally. - New page: 2026 comparison of every AI tool for Excel — 12 tools side by side with honest pros/cons. - New page: Founder bio with full academic and industry credentials — PhD in Statistical Machine Learning, Cambridge, Oxford, UCL Associate Professor, J.P. Morgan VP for AI Strategy. - New page: Press & media kit for journalists and bloggers. - New tutorial: Holt-Winters forecasting in Excel — step-by-step with no add-ins required. - SEO: 41 pages now ship with author Person schema (PhD, alumniOf, affiliation, hasCredential, Royal Academy award) and per-page custom OG images. - SEO: 8 new mid-market competitor comparison pages (vs Databox, Geckoboard, Mode, Metabase, Causal, Rows, Numerous AI, Sigma). 7 May 2026 New ## Free anomaly detector, vertical landing pages, and self-hosted Inter font - New tool: In-browser anomaly detector — rolling z-score on any pasted column. - New pages: 8 vertical landing pages live (e-commerce, SaaS founders, marketing agencies, sales teams, product teams, real estate, operations, HR teams). - Performance: self-hosted Inter font — LCP improved from ~2.9s to ~1.4s. - SEO: 15 new feature and audience pages with FAQPage schema. 5 May 2026 Improved ## Site launched on datahubpro.co.uk — analytics platform live - Launch: DataHub Pro is live at datahubpro.co.uk. - Product: 50+ analytics tools shipping — KPI dashboards, forecasting, cohort retention, RFM segmentation, anomaly detection, AI Q&A with auditable pandas operations. - Pricing: £19/user/month, flat. 14-day free trial. No annual contract. - Demo: Interactive 2-minute demo live. ---