Matching Methods & Rules
Decide what counts as a match. Start from proven defaults and tune matching to your data - by match type, by threshold, or field by field.
Match types
A match type is the starting point: it tells Match IQ what kind of duplicate you're looking for and which data to compare. Match IQ includes proven defaults for the most common goals:
| Match type | Compares | Result |
|---|---|---|
| Individual | First name, last name, address | One record per person |
| Family | Last name, address | One record per family |
| Resident / Household | Address | One record per residence |
| Firm | Firm name, address | One record per company |
| Firm-Individual | First name, last name, firm, address | One record per person per company |
These map directly to real business questions - one mail piece per household, one contact per company, and so on - and each ships with default rules you can use as-is or refine.
Matching methods
Match IQ offers a progression of matching methods, from simplest to most precise:
- Automatic matching - pick a match type and a threshold, and Match IQ selects the appropriate fields and compares them. No field-by-field setup required.
- Rule-based (extended) matching - define exactly which fields to compare and how, prioritizing match fields and deciding match-or-no-match on a per-field basis for full control.
- Advanced matching - find several levels of matches in a single pass (for example, families and the individuals within them) and build associations across data sets, tagging each level with its own identifier.
Thresholds & scoring
Rather than demanding perfect, character-for-character equality, Match IQ scores how similar two fields are. Automatic matching uses four increasingly lenient thresholds - exact, tight, medium, and loose - to decide how alike fields must be to count as a duplicate. An exact threshold requires a 100% match; the others progressively tolerate more variation.
Rule-based matching goes further with weighted scoring: each field contributes to an overall match score according to the weight you assign, so the fields that matter most carry the most influence. You can also force a match or force a no-match when specific conditions are met.
Intelligent comparison
Match IQ's comparisons are built to recognize the countless ways the same information gets entered differently. Depending on how you configure a field, it can account for:
- Transposed letters and minor typos.
- Initials versus full names (for example, J vs. John).
- Abbreviations and substrings.
- Nicknames and name variations, hyphenated and maiden names.
- Blank fields - decide whether a missing value should be ignored, penalized, or evaluated.
Multi-level & associative matching
Real data rarely fits a single level. With multi-level matching you can, in one pass, identify individuals, the households they belong to, and the residences they share - assigning a distinct identifier at each level. Associative matching links related records across different data sets, such as connecting the same person to the different addresses they use at different times of year.