Guides
Choosing a run type
What each of the seven runs answers, and which one to reach for depending on the question you actually have.
There are seven kinds of run. They are not tiers of the same thing — each answers a different question, and picking the wrong one is the main way to spend money for nothing.
Site Explorer
What is this site already ranking for?
The first run on any client. Deduplicated keyword count, position bands, estimated traffic, the full ranked keyword table, and page-level traffic underneath it.
Start here. Everything else is easier once you know the shape of what already exists.
Keyword Research
What else could it rank for?
Expands a seed into everything adjacent, classifies intent, and groups the result into named clusters. The terms that were cut are shown alongside the ones kept, with the reason.
Runs with or without a client attached, so it works for prospecting too.
Content Gap
What do the people beating us have that we do not?
Takes competitors who genuinely rank for your terms and works out what their pages carry that yours does not — format, depth, questions answered.
Technical Audit
What is broken?
Assembles the URL set first, then inspects it, so it audits the pages that matter rather than whatever a crawler happened to reach. Every issue carries a severity and the evidence behind it, and anything you dismiss stays dismissed.
AI Visibility
How often do the assistants mention us at all?
An index across answers already collected in a market. Engine coverage is established before anything is counted, so you always know what the number was counted across.
AI Research
What does a model actually say when asked?
You put your own prompts to the models and the answer is kept as a dated observation.
This is deliberately separate from AI Visibility and always will be. Visibility is a statistic over a corpus of answers; this is one answer at one moment. A model asked twice does not answer twice the same. Showing an index count beside a live citation as though they measured the same thing would be wrong in the most confident-looking way available.
Local and SERP features
What does the results page actually look like?
For a local business the map pack is the bottom of the funnel. Position 41 with a local pack above it is a different problem from position 41 without one, and a report that ignores the layout is describing a different search to the one customers see.