FuzzyMatch

Match two spreadsheets, even when the data is messy

Find the same customers, vendors or contacts across two files despite typos, nicknames and different column names. Free for files up to 500 rows, no sign-up.

How it works

  1. Step 1

    Add two files

    Drop in two CSV, Excel or JSON files: say, a CRM export and a list of event sign-ups. Files are read in your browser first, so you see rows and columns before anything is sent.

  2. Step 2

    Confirm the columns

    FuzzyMatch pairs columns that hold the same kind of data, even when the headers differ (Name vs Full name, Company vs Org). Adjust any pair and choose how strict to be.

  3. Step 3

    Review and download

    Every match gets a score, so you can check the borderline ones and trust the obvious ones. Download the result as a CSV with both records side by side.

What people use it for

Deduplicate a CRM

Find the same customer entered twice with a typo, a nickname or a new email domain.

Reconcile vendor lists

Match supplier names across accounting and procurement systems that spell them differently.

Merge mailing lists

Combine sign-up sheets and newsletter exports without sending people the same email twice.

Clean up research data

Link survey responses to participant records when names and addresses don't line up exactly.

Questions

What is fuzzy matching?

Fuzzy matching finds records that refer to the same thing even when they aren't written identically: "Jon Smith" and "John Smith", "Acme Inc" and "ACME Incorporated". Instead of requiring an exact match, it scores how similar two values are and treats close-enough pairs as matches.

Do I need an account?

No. Guests can match files of up to 500 rows, 3 times a day. A free account raises that to 2,500 rows per file and keeps your results for 7 days.

What happens to my data?

Guest files are processed in an isolated frame that loads no ads or trackers, kept for at most 1 hour so you can download the result, then deleted automatically. They are never written to disk on our servers and never used to train models.

Which file types work?

CSV, TSV, Excel (.xlsx) and JSON arrays of objects. The first row of a spreadsheet should hold the column names.

How is the match score calculated?

Each pair of columns is compared with a method suited to its contents (names, emails, phone numbers, addresses or free text), then combined using the importance you set for each column. A score of 90% or more usually means the same record.