Guide · Diagnosis
A useful audit does not bury a founder under warnings. It turns evidence into a small number of business priorities.
Field data reveals how actual visitors experience speed and stability. Lab tests help diagnose causes but should not be presented as customer behavior.
A technical score alone cannot show whether the right people find the business. Search intent, local results and credible competitors add context.
Every finding should explain what it affects, what to do and why it belongs now rather than later.
Lighthouse runs a controlled laboratory test; CrUX summarizes eligible real-user field data. A useful audit labels both and never presents a lab score as customer behavior. It also records device, date, location and data availability.
The next step is interpretation: which templates are affected, what users were trying to do and whether the issue blocks understanding, trust or action.
Crawling, canonical tags, structured data and indexing matter, but visibility also depends on intent coverage, local signals, reputation and competitors. A brand ranking first for its own name may still be absent from service-led discovery.
Conversion review should follow the actual journey from entry page to proof, CTA, form or WhatsApp and into follow-up. Otherwise the audit stops exactly where commercial friction begins.
Every recommendation should state evidence, likely impact, effort, dependency and owner. A 30/60/90 sequence is more useful than one hundred warnings with equal weight.
Sources can be incomplete and search results change. A responsible audit dates observations, marks illustrative estimates, avoids ranking guarantees and explains what requires further access or specialist review.
Decision checklist