A scan is how a data quality suite checks your data against the rules you've assigned to it. Every scan produces a score you can trend over time, breaks that score down by quality dimension and by object, and lets you drill all the way down to the individual records that failed. This page walks through starting a scan, reading the trend chart, understanding dimension scores, and finding the records behind a result.
Overview
Each suite has a Scans tab where you trigger new scans and review past ones. A scan evaluates every assigned rule against every record in the suite's assigned objects, then rolls the results up into an overall score and a set of dimension scores. From there you can open any scan to see how each object performed, open any object to see how each rule performed, and open any rule to see the exact records it evaluated.
Running a scan
Click Start Scan in the suite header to kick off a run at any time. If the suite has never been scanned, the Scans tab shows an empty state instead — click Start Your First Scan to trigger your first assessment. Once a scan starts, you'll see a confirmation that the scan has begun and the data is refreshing; the scan runs in the background, checking every assigned rule against every record in every assigned object, and the history table updates automatically when it finishes. If you need to stop a run in progress, click Stop Scan.
Scans also run automatically on the suite's configured schedule, so you don't have to trigger every run by hand.
Reading the trend chart and scan history
The top of the Scans tab shows a Scan Performance Trend chart, with headline totals for Total Scans and Avg. Score. Use it to see at a glance whether data quality is improving, holding steady, or slipping over time.
Below the chart, the scan history table lists every run with its execution time, status, the number of evaluation rules and objects involved, how long the scan took, and its overall DQ score. A scan's status is one of Running, Completed, Failed, or Scheduled. Click any row to open that scan's detailed results.
Understanding quality dimension scores
Every record a rule checks gets an individual result — for example Valid or Invalid, Complete or Not Complete — and the rule contributes its configured score toward the overall total. Open a scan to see the Overall Scan Score for that run (it shows as calculating while the scan is still running), along with an average score for each quality dimension:
Validity — whether values conform to the format or constraints a rule expects.
Uniqueness — whether values are duplicated when they shouldn't be.
Completeness — whether required values are present.
Accuracy — whether values correctly reflect what they're meant to represent.
Timeliness — whether data is current enough for its intended use.
Below these headline stats, Object Level Analysis breaks the same scores down per object, so you can quickly see which dataset is pulling the overall score down.
Drilling into an object and its rules
Open any object from a scan's results to see its Rule Level Analysis — one row per rule applied to that object, showing the rule's quality dimension, which of the object's fields it was applied to, any filters that scoped which records it evaluated, and a summary of its result.
Open a rule from that list to see two tabs. Result Analysis shows the rule set's score breakdown, a count of results at the record level, and the filters that were applied along with how many of the object's records they matched. Rule Logic shows how the rule is configured — its attributes, parameters, and a preview of its underlying logic — so you can confirm it's checking what you expect.
Viewing the records behind a result
To see exactly which records drove a rule's result, click View Records from that rule's result. This opens a record-level list for the rule, scoped to the object it ran against, showing every evaluated record and its individual outcome. Instead of guessing what's behind an Invalid or Not Complete percentage, you can inspect the exact rows involved.
Notes
A record can show as Skipped instead of pass or fail. This means the record never reached the rule's logic because it didn't match the rule's pre-filter conditions — skipped records are reported separately and aren't counted toward the pass or fail totals.
If a large share of records show as Skipped, revisit the filter conditions attached to the rule; an overly narrow filter can hide records that should have been evaluated.
The Overall Scan Score and dimension scores only reflect rules that were assigned and actually ran — unassigned rules don't affect the score.
A scan in progress shows its score as calculating until it completes, so wait for a scan to finish before comparing its score against previous runs.