A useful first scan ends with one clear next action. Start with a public page you own or are authorized to test, record what the scanner actually observed, and choose a finding to investigate. You do not need to solve every warning before the report becomes useful.
Choose a page with a real job
Use the page a visitor needs to complete a task: a service page, a product page, or the page that explains how to contact you. The homepage alone may miss problems in other templates.
For a first pass, use an ordinary public URL without a login token, private document identifier, or personal information in its query string. A public-page check does not need access to your customer records.
Write down the URL and the visitor's job before you open the free scan. That gives you context when a report contains several different warnings.
Read the result in this order
- Check the attempt. Did the scan finish, and is its timestamp current? A failed attempt is an observation about the scan, not proof that the website is healthy or down for everyone.
- Check coverage. Look for unavailable categories or individual checks. Missing performance data is not a passing performance measurement.
- Open the finding. Read the evidence, affected URL, and explanation. The grade is a starting point for investigation.
- Choose an owner and next action. Decide who can confirm the issue and make the change.
Performance measurements need extra context. Google's PageSpeed Insights documentation distinguishes a controlled lab run from aggregated visitor data. The two answer different questions; neither should silently stand in for the other.
Turn a warning into a small work item
Here is an illustrative work item you can adapt:
Page: the public services page. Finding: the displayed page title is generic. Next action: the content owner checks whether a more specific title reflects the page. Verification: load the published page again and confirm the intended title is present. Limitation: a title check does not establish search rankings.
Keep the change small enough to verify. If the problem requires access to hosting, application code, or a vendor dashboard, hand the evidence to the person with that access rather than guessing from the score.
Recheck the same scope
After the change reaches the public page, check that URL again. Use the same profile and compare the same finding. Save the new observation alongside the original one.
A successful recheck can support the specific correction it measured. A timeout cannot confirm recovery. A different page, device mode, or test profile may be useful investigation, but it is not a like-for-like comparison.
Decide whether a snapshot is enough
The free scan is a point-in-time check. It is useful for understanding a page and investigating a change. Recurring portfolio work adds a different job: keeping sites and evidence organized between reviews.
If you maintain several sites, walk through the Portfolio example. It uses fictional data to show the difference between a reviewed finding, an incomplete recheck, and verified recovery. For the wider routine, use the agency review checklist.