Share of Model: the metric replacing rankings
Ranking #1 for a query that never generates a click is not a win. The answer engines already decide what most searchers ever see — and here's how to measure whether your brand is in the answer.
Google's own AI Overviews now cover a majority of informational queries. ChatGPT, Gemini, and Perplexity route a growing share of the reader's research through a summary they synthesise before the click ever happens. The blue-link scoreboard is still real — but it measures a shrinking surface.
The metric we actually track
Share of Model asks a simpler question: when the reader queries the category, how often is your brand named in the answer? Across which prompts? Which models? At what position — mentioned first, mentioned in passing, or not at all?
Instrumented well, it becomes a weekly dashboard: prompts monitored, mentions counted, position tracked, deltas against the last cycle, contextualised by which of your published assets the model appears to be quoting.
Why it changes the SEO practice
Traditional ranking work rewards technical hygiene, content depth, and the domain graph. Share-of-Model work rewards citable structure, entity-clear attribution, and the reference trail across the open web the models trained on. The overlap is real, but not total — and most agencies still bill only for the first half.