Testing Data Logic with Test Matching and Near-Miss Detection

Test Matching requires both input mapping and output mapping to be fully configured before it can evaluate a record. If you click Test with output mapping missing or incomplete, it returns a message asking you to complete it first.

1. Navigate to Data Integrity > Data Logic.
2. Complete 1. Pick Blueprint.
3. Complete Step 2. Input Mapping and Step 3. Output Mapping. 

4. Click Test from either mapping step to evaluate a record against your full configuration.

   

   The image above shows 3. Output Mapping of the Data Logic module for a Job Title Normalization and Persona Assignment template, with four Blueprint columns mapped to CRM fields— Normalized Title (Always) mapped to Job Title, Persona (Always) mapped to Persona, Seniority (Output Non-Empty) mapped to Seniority, and Department (Field Empty) mapped to Department — with a teal arrow callout pointing to the Test button at the bottom of the step.

5. Search for and select the record you want to evaluate. If you already know a specific record isn't updating as expected, testing that record directly is the fastest way to see why.
6. Click Evaluate.

Test Matching returns a full breakdown covering both sides of your Data Logic configuration:

• Input Fields — for each input column, whether the record's value matched a condition in the Blueprint, and if not, why. This includes specific details like which Blueprint values were checked, whether a numeric or date value couldn't be parsed, or whether a Blueprint cell was blank.
• Output Fields — for each output column, whether it would be updated, left alone, or filled with a fallback value, and why.

Each field row shows a green checkmark or red X in the Evaluation column, indicating a match or no match, alongside an explanation in the Reason column. An Input Analysis and Output Analysis summary at the bottom of the results gives an overall takeaway for each side of the configuration, so you don't have to piece it together from the individual rows.

If every input field matches something in the Blueprint individually, but the record still doesn't match any row overall, Test Matching flags that too — it means no single Blueprint row covers this exact combination of values, and a new row is likely needed.

You can evaluate as many records as you like in the same session without closing the dialog — select a new record and click Evaluate again.

The image above shows the Test Matching Criteria and Update Conditions dialog in the Data Logic module after evaluating the record "Graso," displaying three input field results — Country/Region and State/Region passing with green checkmarks, and Industry failing with a red X because the field is blank — and three output field results — Country/Region failing because no Blueprint match was found, and Tier and County passing with updates ready to write — with an Input Analysis summary noting the unmatched Industry field and an Output Analysis summary confirming that 2 of 3 output fields would be updated.

When a record doesn't match a Blueprint row, the underlying values are often not actually different — a stray space, a hyphen used in place of a space, or a slash formatted differently than expected. These differences are easy to type past and difficult to catch by eye.

Test Matching detects this automatically. When an input value is nearly identical to a value already in your Blueprint, the diagnostic message calls it out directly and tells you exactly which Match Option would resolve it:

• Whitespace differences — "Business/Commercial" vs. "Business / Commercial." Flagged with a recommendation to enable a whitespace Match Option in Input Mapping. The image above shows a failed input evaluation result in the Test Matching Criteria and Update Conditions dialog for the Type field (Column: Account Type (Exact)), with a teal arrow pointing to the near-miss detection message suggesting that enabling the "Ignore Whitespace" match option would cause the record's value "Business / Commercial" to match the Blueprint value "Business/Commercial," which differ only in whitespace.
• Punctuation and special character differences — "AT&T" vs. "ATT." Flagged with a recommendation to enable Ignore Special Characters. The image above shows a failed input evaluation result in the Test Matching Criteria and Update Conditions dialog for the Company Name field (Column: Company (exact)), with a teal arrow pointing to the near-miss detection message suggesting that enabling the "Ignore Special Chars" match option would cause the record's value "ATT" to match the Blueprint value "AT&T," which differ only in punctuation and special characters.
• Case differences — "Church/Parish" vs. "church/parish." Flagged with a recommendation to enable Case Insensitive. The image above shows a failed input evaluation result in the Test Matching Criteria and Update Conditions dialog for the Type field (Column: Account Type (Exact)), with a teal arrow pointing to the near-miss detection message suggesting that enabling the "Case Insensitive" match option would allow the record's value "church/parish" to match the Blueprint value "Church/Parish," which differ only in letter case.

• Combined differences — some near misses need more than one option to resolve, and Test Matching detects these combinations too:

  • "Other Faith Based" vs. "other faith-based," where a hyphen stands in for a space and the capitalization differs. No single option reveals this match on its own, so Test Matching identifies the full combination needed — Ignore Whitespace, Ignore Special Characters, and Case Insensitive together — in whatever mix actually closes the gap.
  • "New York City" vs. "New York-City," where a hyphen stands in for a space. Test Matching identifies the combination needed — Ignore Whitespace and Ignore Special Characters — in whatever mix actually closes the gap.

  The image above shows a failed input evaluation result in the Test Matching Criteria and Update Conditions dialog for the City field (Column: City (Exact Match)), with a teal arrow pointing to the near-miss detection message suggesting that enabling both "Ignore Whitespace" and "Ignore Special Chars" match options together would allow the record's value "New York-City" to match the Blueprint value "New York City," which differ in how words are separated. 

Test Matching is built for checking one record at a time in detail — but before committing changes across your entire dataset, it's worth running the template in Preview mode first. Preview evaluates every record the template would affect and shows exactly what would change, without writing anything to your CRM.

Where Test Matching explains why a single record matched or didn't, Preview shows what would happen across every record at once — the before and after value for each output field, for every record in scope. This is especially useful once you've used Test Matching to confirm your logic works on a handful of individual cases and want to see the full-population effect before running the template for real.

Preview is also the tool used to troubleshoot drift after a Blueprint update — since it shows exact before-and-after values per record without changing any data, it's a better fit for understanding why a value changed than any aggregate report could be.

To run Preview:

1. Complete 1. Pick Blueprint, 2. Input Mapping, and 3. Output Mapping.
2. Optionally, focus the records the operation looks at under 4. Filter Records.
3. Under Run, click Review.

4. Select Preview as the run mode before running the template. 

   

   The image above shows the Mode step (1) of the Data Logic run dialog, with Preview selected — indicating no data will be modified — and Update available as an alternative option that would write changes to the CRM.

In the emailed preview CSV file, review the before/after values shown for each affected record.   Unmodified records will not appear in the CSV — only records the template would actually change are included. 

The image above shows a Preview mode CSV export from the Data Logic module, displaying input fields used for matching — Industry, Employees, and Annual Revenue — alongside Before and After values for three output fields — ICP Score, ICP Tier, and ICP Flag — for each evaluated record, allowing review of what changes would be written to the CRM before running in Update mode.

Once you're confident in the results, switch to Update mode to apply the changes.