FuzzyMatchMatching & confidence scores
After you add files, FuzzyMatch pairs columns with similar names and suggests pairs based on the data in them. You stay in control: every column in the first file has a Compare with menu and an Importance setting.
Pairing columns
Pair columns that hold the same kind of value:
| First file | Second file | Pair? |
|---|---|---|
| Full Name | Contact Name | Yes |
| Email Address | Yes | |
| Company | Organization | Yes |
| Customer ID | Account No | No. IDs from different systems never match and only lower the score. |
Choose Don't compare for any column that would add noise. You need at least one pair to run a match.
Importance
Importance sets how much each pair counts toward the overall score.
- High: columns that identify a record on their own, like email or full name.
- Medium: the default. Useful but not decisive, like company name.
- Low: columns that often differ for the same record, like city, or phone numbers stored in different formats.
A useful starting point for contact lists: name and email on High, company on Medium, city and state on Low.
What the score means
Each match gets a confidence score from 0 to 100%. FuzzyMatch compares every paired column, then combines those comparisons using the importance you set. 100% means every compared value matched.
| Score | Colour | What to do |
|---|---|---|
| 90% and up | Green | Almost certainly the same record. |
| 75–89% | Blue | Very likely the same; spot-check a few. |
| Below 75% | Amber | Worth a look: could be a nickname, a typo, or a different record. |
How strict?
Strictness is the lowest score that still counts as a match.
- Loose (60%) catches heavy typos and nicknames, but you'll review more false matches.
- Balanced (75%) is a good default for names and companies.
- Strict (90%) keeps only near-identical records. Use it when a wrong match is costly.
If you get too few matches, loosen the setting or pair more columns. If you get too many wrong ones, tighten it or raise the importance of an identifying column like email.
Matches per record
- Best match only pairs each record in the first file with its single closest match.
- All good matches lists up to 10 matches per record that clear your strictness. Use it to find duplicates or to see the runner-up candidates.