Common setup mistakes

Use the public brand name customers would recognise, a search term that identifies the brand, and the canonical website domain. Avoid internal project names, campaign names, or domains that redirect somewhere unrelated.

The brand domain is important for citation and page matching, so check spelling, regional variants, and subdomain choices before running the first audit.

Very broad terms create noisy reports. Start with customer discovery, comparison, and buying-intent terms that the brand should realistically appear for.

Use prompts for natural-language questions that customers ask AI systems. The audit combines keywords and prompts into the set of terms to test, so duplicates waste limited tracking slots.

Add competitors that customers would actually compare against. Competitor benchmarking, share of voice, citation gaps, and recommendations are more useful when competitors have accurate names and domains.

If a brand has no competitors, the product can still audit visibility, but competitor comparison and market share context will be limited.

The first audit is a baseline. Trend labels, period comparisons, resolved recommendations, and regression checks become more useful after additional audits with stable inputs.

Avoid changing keywords, prompts, competitors, market, or domain immediately after the first audit unless the setup was wrong. Stable inputs make movement easier to interpret.

Technical Review expects public HTTP or HTTPS pages. Localhost, private network addresses, unsupported schemes, oversized batches, or pages that block crawling can prevent useful results.

Scan the page most relevant to the keyword or opportunity you want to improve, then re-scan after changes go live to create a fresh PageAudit snapshot.