Action Plan

Before any project is raised, each question is classified: whether it names your business and what it asks for. Success follows from that: a request for your website or contact details succeeds when the answer gives the correct ones, a factual question when the answer is accurate and supported, and only a discovery or comparison question when the answer puts you forward. A project that assumes a recommendation was needed is not raised from a question that did not call for one, and a topic read from a URL path alone (such as /new/ or /practice/) is not raised at all. A project stays in Investigate until a person supports its diagnosis with evidence, and, for a content gap, records your passage and the competitor's or answer's passage side by side with the question intent; only then is it Ready to implement.

Every recommendation is one project in a shared queue: the run plan, the competitor cards and the stored recommendations each feed it, and the same work found twice is one project. Projects that share a question or a target are grouped as one initiative, and each keeps its own status. Each project says what was observed, whether the cause has been reviewed, who owns it and where it stands. Doing the work and seeing an effect are recorded separately: an effect is read only from a retest that asks the same questions. The plan at the top comes from the answers in your latest measurement run. Every audit also produces a list of specific gaps: places where your brand is missing, cited wrongly, or beaten by competitors. Those are grouped by where you have the most control (use the section list to move between groups) and ordered within each group by how much work they take, lightest first. Every recommendation shows the prompts it came from, so you can read the answers behind it before acting.