Match summary & profile review
Audience: QA, marketer, and data or analytics engineer
Prerequisites: IDR overview →, Setup steps →, Prepare your data →, Golden Record →, Deterministic →, Probabilistic →
Inspect your identity graph before syncing any data to make sure everything is matching as expected.
Why it matters
The Summary and Profiles tabs give you visibility into how records are matching. Use these tools to:
-
Validate that match rates look reasonable
-
Catch unexpected merges or suspicious profiles
-
QA survivorship rules
-
Understand confidence levels before syncing
Use Cases
| Use Case | What You Can Validate |
|---|---|
| New model configuration | Understand dedupe rate and identifier coverage |
| Probabilistic rollout | Compare match behavior before and after enabling |
| Golden Record QA | Ensure canonical traits look correct |
| Ongoing monitoring | Track match rate changes over time |
Summary tab overview
The Summary tab gives you a high-level dashboard view of model performance:
-
Total number of identities (
HT_IDs) created -
Overall match rate (% of records grouped into identities)
-
Breakdown by confidence level (Exact, Strict, Loose)
-
Identifier coverage (% of rows with usable values)
-
Match rate trend over time (if the model has run multiple times)
Use this view to gauge how effective your model is at unifying records.
Profiles tab overview
The Profiles tab lets you inspect specific resolved identities. Each profile includes:
| Attribute | Description |
|---|---|
HT_ID | Unique identity ID |
| Source records | All the rows merged into this profile |
| Confidence level | Match score (if probabilistic) |
| Used identifiers | Which fields were used to link records |
| Golden Record traits | The values selected for each canonical field |
You can search by email, phone, or ID to inspect specific users.
What to look for
| QA Task | What to check |
|---|---|
| Identifier gaps | Records missing emails or names won’t match—check identifier coverage |
| Unexpected merges | Use Profiles to debug if records were merged incorrectly |
| Golden Record values | Confirm that canonical fields look clean and trustworthy |
Best practices
Do
-
Check dedupe rates for each identifier type (email, name, phone)
-
Check dedupe rates by model
-
Use the Array survivorship rule for fields like emails if you want full visibility
Don’t
- Sync records without reviewing Summary and Profiles first
Before you sync
Use this checklist:
-
Match rate looks healthy based on your goals
-
Spot-check Profiles for key users
-
Golden Record traits look correct
-
Required fields are present in the output table
-
Confidence tiers are configured correctly for each use case
What’s next?
Ready to build your audiences?
- Build audiences using clean, trusted traits