Outrings
Accessibility, mobile and presentation

How often should I audit my website?

After every significant deployment, and on a schedule between them. Configuration regresses quietly, and almost always through a change nobody thought was risky.

3 min read
Short answer

After every deployment that touches templates, headers, robots.txt or hosting configuration, plus a scheduled check monthly. The scheduled one exists because the damaging regressions are invisible — a noindex tag, a blocked crawler, an expired certificate — and nobody reports them.

Why scheduled checking matters

The faults that cost the most are silent. A site with a stray noindex looks perfect to every visitor and to everyone on the team. It disappears from search over weeks, and the first signal is usually a traffic report a month later.

  • A noindex deployed from staging.
  • A robots.txt replaced by a platform default.
  • A certificate whose auto-renewal stopped.
  • A canonical tag pointing at staging after a migration.
  • A CDN rule blocking crawlers, added for unrelated reasons.
  • A third-party script that started failing.

None of these produce a complaint. All of them are visible in a check that takes seconds.

A workable cadence

WhenWhat to check
Every deployment touching templates or configTitles, canonicals, noindex, structured data, headers
WeeklyUptime, certificate expiry, the homepage and top landing pages
MonthlyA full audit: crawlability, security headers, broken links, accessibility
QuarterlyThe wider picture — content still accurate, dates still honest, dead pages removed
After any migrationEverything, twice — before and after DNS moves

Compare, do not just measure

A single score has limited value. What is useful is the difference between two audits: which findings are new, which were resolved, which category moved. That converts an audit from a report into a signal.

This is what a snapshot history gives you — audit, store the result, and the next audit tells you what changed rather than only what is true. A regression introduced last Tuesday shows up as a specific check that started failing, with a date attached.

Automate the routine part. A scheduled job that re-audits your key pages and alerts you when a score drops or a new failure appears turns this from a discipline you have to remember into infrastructure you can forget about. Manual monthly audits get skipped in busy months, which is exactly when deployments are riskiest.

What to do with the results

  1. Fix anything that removes you from an index — noindex, robots blocks, canonical errors. Same day.
  2. Fix security exposures — leaked credentials, readable sensitive paths. Same day.
  3. Fix accessibility barriers that block use entirely — keyboard traps, missing labels.
  4. Schedule the rest by effort against impact.
  5. Re-audit after fixing, and confirm the finding is actually gone rather than assuming.

That last step catches more than expected. A fix that was deployed to the wrong environment, or a change reverted by a subsequent merge, looks identical to a fix that worked — until you check.

What our audit reports about this

Every item below is measured directly, not inferred. Run it against your own site and the result names the exact rule or header responsible.

  • A full audit across 18 categories, with every finding dated.
  • Stored snapshots per site, so successive audits can be compared.
  • A changes endpoint reporting new problems, resolved findings and category movement between snapshots.
  • Scheduled re-auditing, so history accumulates without anyone remembering to run it.

For agents and scripts, the same measurement is at /api/v1/infra?url=yoursite.com — see the API documentation.

Related questions

Is a monthly audit enough?

For a stable site, usually. For one deploying weekly, check after each deployment — that is when things break, and a monthly cadence means up to four weeks of invisible damage.

What should I check after a migration?

Everything. Migrations break redirects, canonicals, robots.txt, certificates and structured data simultaneously. Audit before the DNS change, immediately after, and again a week later once caches have turned over.

Can I automate this?

Yes, and you should. A scheduled re-audit with alerting on regressions removes the failure mode where the check gets skipped precisely when it is most needed.

Read next

All 50 guides · How every check works · API for agents