Did the last deploy fix it, or move it?
Every run is compared against the one before it, issue by issue rather than total against total. You get what you fixed, what you introduced, and what has been sitting there unchanged.
Scan a public URL free · no credit card required
color-contrastFixedimage-altIntroduced- Vs the previous run
- issue by issue, not total vs total
- Fixed · Introduced · Unchanged
- three numbers on the report
- By issue identity
- rule, page and element fingerprint
- Free to start
- scan a public URL, no card
Did the last deploy help? The next report says so
No separate comparison to run. The delta sits on the report you already opened.
- Step 1
Run the scan
Scan your URLs once to establish the baseline run — the report you will compare the next one against.
- Step 2
Ship, then re-run the same scan
Re-run the scan you already have rather than creating a new one — runs belong to a scan, numbered in order. A schedule does this for you on a daily, weekly or monthly cadence.
- Step 3
Read the delta on the report
The report compares against the run immediately before it and shows three numbers — fixed, introduced and unchanged — plus the number of the run it compared against.
Three numbers, on the report you already opened
Compared by issue identity — the rule, the page and the element — never by totals.
Fixed
A fingerprint that was present in the previous run and is absent now — evidence that the change landed.
Introduced
Present now and absent before. The newest code, the cheapest fix, and the number a falling total hides.
Unchanged
Present in both runs — the backlog you have had the opportunity to clear and have not.
By identity, not totals
Each finding is fingerprinted by the rule that fired, the page URL and the element selector. It is set arithmetic on identity, not a comparison of totals.
On the report you already opened
The comparison used to be reachable only from a separate compare-two-scans screen. The same comparison now sits on the report itself.
Compare arbitrary runs too
The compare screen still exists to line up any two runs. What sits on the report is specifically the comparison against the previous run.
Frequently asked questions
The run you are looking at, against the run immediately before it in the same scan. You get three numbers — fixed, introduced, unchanged — plus the number of the run being compared against, so it is always clear what the baseline was.
Each finding is fingerprinted by three things: the rule that fired, the page URL it fired on, and the element selector it fired against. An issue present before and absent now is fixed. Absent before and present now is introduced. Present in both is unchanged. It is set arithmetic on identity, not a comparison of totals.
Partly, and it is worth knowing where it stops. If a fault moves to a different selector or the URL changes, the fingerprint changes, so the same underlying problem reads as one fixed and one introduced. The numbers stay honest about what was observed; they are not doing semantic matching of your markup.
No. The report compares against the immediately preceding run. There is no chart of scores over months and no score history to export. Criteria no automated engine can judge never appear as fixed or introduced, because they were never measured in either run. And nothing alerts you — the numbers are on the report when you open it. Nothing pages you because a run introduced twelve issues.
Re-run the scan you already have rather than creating a new one — runs belong to a scan, numbered in order. A schedule does this for you on a daily, weekly or monthly cadence. A comparison against the previous run is only as good as the previous run: the more recently the last one ran, the more a regression is attributable to a single deploy.
Yes, the compare screen still exists for that. What sits on the report is specifically the comparison against the previous run, because that is the one that answers 'did what I just did work'.
Related
- Scheduled ScansDaily, weekly or monthly runs in your own timezone
- WCAG 2.2 CoverageWhat automated rules reach, and what is marked not checked
- WCAG 2.2 ChecklistAll 55 A and AA criteria, and which a machine can reach
- All of AccessibilityAutomated WCAG scanning, workflow audits, and reports
- PricingPer-seat test management and per-thread device testing, priced on the page
Find out what your last release introduced
Scan a URL, ship, scan it again. The second report tells you which way it went.