Map Extract
Free tool

Find near-duplicate company and people's names

Paste a list of names and the de-duplicator groups the ones that are probably the same, even with Ltd added, & written as and, a missing space or a typo. You choose which spelling to keep in each group, then copy the de-duplicated list or download a report of every group.

Runs in your browser. Nothing you enter is uploaded.

Eleven names, four businesses

This is a merged list from three sources with the usual variations: legal suffixes, The at the front, & and and, a joined word and two misspellings.

At the default 90% threshold the eleven names fall into four groups. Before comparing, names are lower-cased, & is read as and, and words such as the, ltd, limited and co are ignored, which is why the first three pottery names score 100%. In each group the longest spelling is preselected as the one to keep; change it with the radio buttons if you prefer the trading name.

Input
Kiln & Co Pottery
Kiln and Co Pottery Ltd
The Kiln & Co Pottery
Kiln Co Potery
Marlow Print Works Limited
Marlow Print Works
Marlowe Printworks
Oak Lane Joinery
Oaklane Joinery Ltd
Pip & Parsnip Deli
Pip and Parsnip Delicatessen
Result
Group 1: Kiln & Co Pottery 100%, Kiln and Co Pottery Ltd 100% (kept), The Kiln & Co Pottery 100%, Kiln Co Potery 98%
Group 2: Marlow Print Works Limited 100% (kept), Marlow Print Works 100%, Marlowe Printworks 98%
Group 3: Oak Lane Joinery 98%, Oaklane Joinery Ltd 98% (kept)
Group 4: Pip & Parsnip Deli 93%, Pip and Parsnip Delicatessen 93% (kept)

After de-duplication: 4 names

Using the de-duplicator

  1. Paste the names

    One name per line, up to 5,000. Only the names are needed; keep the rest of each record in your spreadsheet.

  2. Say what kind of names they are

    Company names strips legal suffixes and common filler words. People's names removes titles such as Dr, reads Shah, Priya as Priya Shah, and lets P. Shah match Priya Shah.

  3. Set the threshold

    Drag Match threshold between 80% and 99%. Lower values group more loosely and catch more variation, along with more false matches.

  4. Choose the survivor in each group

    Each group lists its members with a similarity score and line number. The radio button marks the version that stays.

  5. Export

    Copy de-duplicated list and Download list give every name not in a group plus your chosen survivors, in the original order. Download groups report lists every group, member, score and whether it was kept.

Review every group

A high score means two names are spelled alike, not that they are the same business. Two different firms can share a name in different towns, and with people's names Tom Reed and Tom Reid score 95% and may well be two people. Compare the town, postcode, website or company number in your full records before you merge anything.

Names are only compared with names that begin with the same letter once cleaned (and, for people, with names sharing a surname), which keeps large lists fast. The note under the groups shows how many pairs were compared. A pair whose cleaned names start with different letters, such as a firm listed under its initials, will not be grouped.

Questions people ask

What is fuzzy matching for company names?

Fuzzy matching scores how similar two names are rather than requiring them to be identical. This tool cleans each name first, removing case, punctuation and words like Ltd and The, then scores pairs with Jaro-Winkler and Levenshtein measures. Names whose score reaches your threshold are grouped as probable duplicates for you to review.

Which threshold should I use?

Start at the default 90%. It catches suffix and spacing differences and small typos while keeping false matches rare. Drop towards 85% if obvious duplicates are being missed, and raise it if unrelated names are being grouped. Always read the groups at lower thresholds before exporting.

Can it de-duplicate a whole CSV, not just names?

It works on a list of names. To clean a full CSV, paste the name column here, decide the groups, then use the groups report to mark or delete the duplicate rows in your spreadsheet. For exact duplicates across whole rows or chosen columns, the CSV deduplicator works on the file directly.

Does it handle people's names written surname first?

Yes, in People's names mode. Shah, Priya is read as Priya Shah, titles such as Dr, Mrs and Prof are ignored, and an initial can match a full first name, so P. Shah, Priya Shah and Dr Priya Shah can end up in one group.

Where your data goes

Names are cleaned and compared on your device; the note under the groups shows how many pairs were checked locally. The list is not uploaded or stored.

Checked against the tool on

Top