Verato MDM Cloud™ assigns a match score every time it compares identity data. The score shows how closely two sets of identity data agree. It also determines whether records are linked automatically, sent to a data steward for review, or kept separate.
How match scores are calculated
Match scores are produced by Verato's matching algorithm, which uses Verato Referential Matching® to compare incoming identity data against Verato's reference database and your existing identities. The same algorithm is used in two places:
- Search. When you search for an identity, each result is compared with your search criteria and receives a score based on how similar its attribute data is to your input.
- Record posting. When you add or update (post) a source record, the score determines whether that record matches an existing identity.
Match score scale
Match scores range from 0.0 to 1.0 and are expressed as decimals (for example, 0.85).
| Score | What it means |
|---|---|
0.0 | No similarity between the input and the identity record. |
Greater than 0.0 and less than 1.0 | Partial similarity. The higher the score, the more closely the input agrees with the identity record. |
1.0 | The highest level of similarity between the input and the identity record. |
How match scores drive matching decisions
Each add or update is evaluated against two thresholds. Together, they decide what happens to the record.
| Score range | Result |
|---|---|
At or above the Automatic Match threshold (default 0.8) | The records are considered a definite match. Verato links them under a single LinkID. |
| Below the Automatic Match threshold and above the Manual Review threshold | The records are considered a potential match. Verato does not link them automatically and creates a data stewardship task so a data steward can decide. |
| Below the Manual Review threshold | The records are not considered a match and are not linked. |
Note: Both thresholds are configured for your instance. To review or change them, contact Verato Support.
Note: Records can also be linked with a score below 0.8. When the incoming and existing records match the same reference identity in Verato's reference database and no conflicts are found, Verato links them regardless of the score. If a task is created for the link, such as a Match – Same Source task, its score compares the existing record with the incoming record plus reference data, so it can fall below the threshold.
Why a Potential Match task can show a score of 0.8 or higher
When you post a record, Verato calculates two scores before deciding whether to link it to an existing identity:
- Augmented score. The incoming record, combined with data from its matching reference identity, is compared with the existing identity.
- Head-to-head score. The incoming record is compared directly with the existing identity, without any reference data.
The task records only the augmented score. If the augmented score is at or above the Automatic Match threshold but one of Verato's safeguards prevents the link, you see a Potential Match task with a score of 0.8 or higher. The head-to-head score is lower, and may include a penalty, but it is not shown on the task.
A Potential Match task with a score of 0.8 or higher is created in the following scenarios.
| Scenario | What happens |
|---|---|
| 1. The records match different reference identities | The incoming and existing records each match a different reference identity in Verato's reference database. This indicates they may represent different people, so Verato does not link them. |
| 2. The records match the same reference identity, but a head-to-head penalty applies | The direct comparison shows a pattern that suggests two different people, such as twins, a parent and child, or other family members. Verato does not link them. |
| 3. The existing identity has a multiple conflicting match | When the existing record was originally posted, it matched more than one reference identity. Because its reference match is ambiguous, Verato can prevent the link and create a task instead. |
| 4. The records were migrated with existing MPI IDs | If your records were loaded from a previous Master Person Index (MPI) with their existing MPI IDs preserved, Verato keeps identities with different MPI IDs separate. When Verato determines that two of these identities are the same person, it creates a Potential Match task so a data steward can review them. |
In scenarios 1–3, the task score reflects the augmented score only. In scenario 4, the high score reflects Verato's match result. The records stay separate only because they carry different MPI IDs.
Match override rules
If your instance uses match override rules, they are evaluated after standard matching. Override rules can adjust match scores, bonuses, or penalties based on your business rules, which can change the final match decision.
Comments
0 comments
Article is closed for comments.