Data cleaning: turning an export into something you can use

Cleaning the file once is worth almost nothing. The job is to make the next export clean itself.

Nearly every spreadsheet problem that looks like a formula problem is a data problem: dates stored as text, three spellings of the same company, trailing whitespace, numbers with currency symbols baked in, merged cells left over from a human-readable report that was never meant to be data.

We do the cleanup, and then we leave behind the routine that did it — a set of helper sheets or an Apps Script pass you can run on next month's export in one click. That is the difference between paying for this once and paying for it every month.

You probably recognise at least one of these

VLOOKUP and MATCH returning #N/A on values that clearly exist
Duplicate customers or products under slightly different names
Dates that sort wrongly because half of them are text
Numbers that will not sum because they carry symbols or non-breaking spaces
A report export with merged cells and repeated header rows

What you get

A cleaned dataset with the original preserved untouched alongside it
A documented list of every transformation applied and every record changed
Fuzzy-matched duplicate report for you to approve before merging
Data validation rules so bad values cannot re-enter by hand
A re-runnable cleaning routine for future exports

How the engagement runs

Profile the data

We run a column-by-column profile — types, blanks, outliers, distinct values — and show you what is actually wrong before quoting.

Fixed quote

Priced on the profile, not on guesswork.

Clean with an audit trail

Every change is logged. Nothing is silently overwritten.

Hand over the routine

So the next file takes you five minutes, not five days.

Questions we get asked

How do you decide which duplicate is the real one?

We do not decide. We produce a match report with a confidence score and the differing fields side by side, and you approve the merges. Automated merging of customer records without review is how data disasters happen.

Can you handle files too big for Sheets?

Yes. Above roughly 5 million cells we will clean outside Sheets and load the result back, or recommend BigQuery if the volume genuinely warrants it.

Is our data confidential?

We will sign an NDA before receiving anything, and we work in your Drive rather than downloading copies wherever that is possible.

Tell us what it needs to do

Share the file or describe the outcome. We reply with scope, price and a delivery date.

Get a fixed quote