Match Management is where you review and act on every record the integration has tried to match. Open it any time you want to check what got matched, what still needs a decision, and what could not be matched at all.

You will find it in the Altrata Settings app, on the Match Management tab.

Match Management groups every eligible record into three buckets. The number next to each tab tells you how many records are currently sitting there.

Every tab also lets you switch between Lead, Contact, and Account (and Person Account, if your instance uses it) using the sub-tabs near the top.

You will see an Auto-match threshold indicator in the top right of every Match Management tab, set to 95% by default.

This is the confidence score a match has to hit before the system links the record automatically. Anything at or above the threshold ends up in Matched. Anything below it ends up in Possible Match for your review.

If you want auto-matching to be more aggressive (more records flowing straight into Matched without manual review), lower the threshold. If you want to be stricter and review more candidates by hand, raise it.

Change the threshold from the Field Mapping tab in Altrata Settings, right next to Match Management.

The Matched tab shows every record that has been successfully linked to an Altrata profile.

Each row shows:

Use this tab when you want to spot-check matches after a bulk match run, audit the quality of recent auto-matches, or confirm that a specific record was linked correctly.

The list is ordered by recently added by default, so the newest matches sit at the top.

The Possible Match tab is where most of your day-to-day match management work happens. These are records where the system found candidates but the confidence score did not clear the auto-match threshold.

Each row shows your Salesforce record on the left, followed by up to three possible matches for individuals and five possible matches for organizations, each with a confidence percentage and the basic profile info (name, role, organization, public/private status, location). On the far right is a Reject option for the whole row.

For each row, do one of three things:

Once you have made your selections, click the confirm [n] selections button in the top right (where [n] is the number of selections you have made). A confirmation modal will pop up asking you to confirm the action across all the records you selected.

This batch confirm is the fastest way to clear a backlog of possible matches. You do not need to confirm row by row.

The No Match tab shows records the system could not link to anything in the Altrata dataset. Each row shows the Salesforce record on the left and a No Match Reason on the right.

Common reasons you will see:

A record in No Match is not necessarily a problem. Sometimes the person or organization genuinely is not in the Altrata dataset, and that is fine. But if you expected a record to match and it did not, here is where to start:

On the No Match tab, you will see a Send Request button on rows where requesting a profile is supported. Click it to redirect to the Altrata web platform, where you can submit a profile request to the research team.

The research team reviews requests and responds through the standard fulfillment process. Once a profile is created and the record is matched in a future match run, it will move out of No Match automatically.

There are a few moments when Match Management is most useful:

Records may also be matched on a case-by-case basis. Should this be enabled, users can manually match Leads, Contacts, Person Accounts, or Accounts.

Manually Unlink a Record

Common questions about how the Altrata matching engine evaluates records, assigns scores and decides whether a record is an auto-match, a possible match or no match.

Are the matching rules evaluated independently or combined?

The matching logic runs all rules at the same time. The best match found from any combination of rules is retained as the match.

Does satisfying a single rule result in an auto-match, or are multiple rules evaluated collectively?

A single rule can result in a match. Whether it becomes an auto-match or a possible match depends on the match score.

How is the overall match score calculated when multiple rules match?

There are different layers of matching. The first layer is an exact match, which scores 100. The next layer is fuzzy matching, where different weights are assigned to different data points. Those weights combine to produce the overall score.

Is there a priority order in which the rules are evaluated?

There is no formal priority order, but the engine always runs exact matching on the data first. Fuzzy matching only runs on records where no exact match is found.

When multiple data points have exact matches:

Does the matching engine use exact matching or fuzzy matching?

Both. Exact matching runs first. If no exact match is found, fuzzy matching runs on the record.

Are spelling variations, abbreviations, or formatting differences considered?

Yes. The engine handles examples like:

The score assigned depends on how close the fuzzy match is. Smaller differences produce higher scores, larger differences produce lower scores.

Does the matching engine recognize nicknames?

Yes. The engine maintains a set of name synonyms it uses during matching. Common examples include:

Work is underway to also ingest alternate names into the matching logic in future releases.

How do nickname matches affect the match score?

When synonyms apply, the engine evaluates the record against every reasonable name combination.

For example, if an Altrata profile exists for "Robert Smith," the engine will attempt to match it against:

Each combination is scored alongside the other data points on the record (organization, email, address, etc.), using the model's defined weights and fuzziness tolerances.

What weight is assigned to Organization Name within the matching algorithm?

Organization Name is weighted differently depending on which object is being matched:

Can a record be auto-matched when the person's name matches but the organization differs significantly?

Yes, as long as another strong data point is present. Examples:

If the only data points on the record are name and organization, and the organization name differs significantly, the record cannot be auto-matched.

Under what circumstances would such a record be classified as an Exact Match rather than a Possible Match?

It will not be classified as an Exact Match unless there is another data point that produces a high-confidence match. Name alone is not enough.

Does address matching require all address components to match?

No. ZIP code alone or a partial address is enough to produce a match. The more fields that overlap, the higher the score.

How are missing address attributes handled?

The engine uses only the address fields that are submitted. There is no imputation or fallback logic for missing fields.

Is partial address matching supported?

Yes.

What exactly constitutes a 100% match?

A 100% match is an exact match on one of the following combinations, with no fuzzy matching applied:

Does the match need to satisfy all of the configured attributes?

No. Any one of the combinations above is enough.

Can records that rely on fuzzy matching still score 100%?

No. Only exact matches can achieve a score of 100.

When Auto-Match Threshold is set to 100%, are records scoring exactly 100% automatically linked?

Yes.

Does any match below 100% appear under Possible Match?

Not quite. Only matches with a score of 61% or higher appear in Possible Match. Anything below 61% is classified as No Match.

Are there additional rules or thresholds that affect this behavior?

No.

Is there a way to see a detailed score breakdown for a matched record?

Not today. The matching model returns an overall score for each match, but it does not expose a per-field breakdown (for example, Name 40%, Organization 30%, Address 20%, Email 10%). A detailed breakdown can be calculated manually by applying the model's defined weights, but that information is not captured at the field level in the app.