Data-quality how-to

How to find and fix errors in Excel data

Most bad reports start with bad data. Excel silently turns part numbers into dates, stores numbers as text, and hides duplicates in plain sight — and 41% of finance teams say catching these errors is their hardest task. Here’s the checklist to find them before you build on top, and how to auto-scan a file in one step.

£28k +18%
1.1–1.5bn
people use spreadsheets worldwide
EarthWeb
~20 hrs
a week knowledge workers spend in spreadsheets
Acuity Training
~94%
of operational spreadsheets contain at least one error
Panko / EuSpRIG
~2 min
from a raw file to an auditable result with DataHub Pro
DataHub Pro

The errors Excel makes silently

The dangerous ones don’t show a #ERROR — they look fine.

Everyone knows the visible errors — #REF!, #DIV/0!, #N/A. The dangerous ones are the silent ones, where the data looks correct but isn’t. The signature example: Excel auto-converts an entry like ‘12-14’ or a code like ‘MARCH3’ into a date, permanently changing the underlying value to a serial number. Once saved, the original is gone.

Just as common: numbers stored as text (so they won’t sum), inconsistent types in a column (mostly numbers with a few ‘N/A’ strings), stray duplicates, and blanks that break aggregations. None of these throw an error — they just quietly produce a wrong total, and you find out in the board meeting.

The fix is a habit: run a data-quality pass on every file before you build dashboards or reports on it. A five-minute check up front saves a retracted number later.

The data-quality checklist

What to look for, and how to catch it in Excel.

CheckHow to spot itHow to fix
Mangled / auto-converted datesNumbers where you expect codes; dates in a text columnFormat column as Text before pasting; re-import raw
Numbers stored as textLeft-aligned numbers; green triangle; SUM ignores themConvert to number (Data → Text to Columns, or *1)
Duplicate rowsData → Remove Duplicates preview; COUNTIF > 1De-duplicate on a key column
Blanks & inconsistent typesGo To Special → Blanks; filter a columnFill, flag, or exclude; standardise type
OutliersSort, or z-score / conditional formattingInvestigate before averaging over them
Broken formulasFormulas → Error Checking; trace precedentsFix references; watch global undo across workbooks

Why validation should come before analysis

Building on unchecked data is the most expensive mistake.

The reason this matters more than it sounds: every downstream artefact — the dashboard, the forecast, the board report — inherits the errors in the source. A single column of text-numbers means a total is silently understated everywhere it appears. Because nothing flagged it, the error propagates with full confidence.

That’s why data validation belongs before analysis, not after someone questions a figure. Catching a type mismatch or a duplicate at upload is cheap; catching it after it’s in a client deck is not.

Auto-scanning a file in one step

Let the tool flag the problems for you.

DataHub Pro runs this checklist automatically the moment you upload. It flags likely date-mangling, columns where numbers are stored as text, duplicate rows, blanks, inconsistent types and statistical outliers — before you build anything — and its cleaning tools fix them in place. You start from data you can trust rather than discovering the problem in the output.

Because the scan is automatic and consistent, you don’t have to remember the whole checklist every time, and you don’t have to be the person who spots that one text-number in a 5,000-row file. Upload, review the flags, fix, then build.

Frequently asked questions

How do I find errors in Excel data?

Run a data-quality pass: check for auto-converted dates, numbers stored as text (left-aligned, green triangle), duplicate rows, blanks, inconsistent types and outliers. Use Error Checking for formula errors. Or upload the file to DataHub Pro, which scans for all of these automatically.

Why does Excel turn my data into dates?

Excel auto-converts entries that look like dates — e.g. ‘12-14’ or codes like ‘MARCH3’ — into date serial numbers, permanently changing the value. Prevent it by formatting the column as Text before pasting, or by importing the raw file without auto-conversion.

How do I fix numbers stored as text in Excel?

Text-numbers are usually left-aligned with a small green triangle and are ignored by SUM. Fix them with Data → Text to Columns, multiplying by 1, or VALUE(). A tool that flags the whole column at once saves checking cell by cell.

How do I find duplicate rows in Excel?

Use Data → Remove Duplicates (preview first) or COUNTIF on a key column to flag rows that appear more than once. De-duplicate on the column that uniquely identifies a record, not the whole row.

Can I check a whole file for errors at once?

Yes. Rather than running each check manually, DataHub Pro scans an uploaded file for mangled dates, text-numbers, duplicates, blanks, inconsistent types and outliers in one pass and lets you fix them before building reports.

Why is data validation important before analysis?

Every dashboard, forecast and report inherits the errors in its source data, and silent errors (like text-numbers) produce wrong totals with no warning. Validating at upload catches problems cheaply, before they reach a client deliverable.

Explore related guides

More guides on cleaning and reporting from spreadsheets.

Guide
AI Excel data cleaning
Fix messy data with AI.
How-to
Compare two Excel files
Spot what changed.
How-to
Variance analysis in Excel
Explain the numbers.
How-to
Excel to dashboard
From clean file to dashboard.
Function
XLOOKUP in Excel
Match and look up data.
Guide
AI for Excel
The complete map.

Catch bad data before it reaches your report

Upload a file and DataHub Pro flags mangled dates, text-numbers, duplicates, blanks and outliers automatically — then helps you fix them. Free tier, then $14.99/mo.

Try it free on your file →