The Recipe Blog That Got Cited Everywhere and Visited Nowhere
A composite case study based on patterns widely reported across recipe and how-to sites. Shows why appearing constantly in AI Overviews does not automatically mean more visitors - and what actually closed the gap.
This is an independent breakdown written by INDEXED. We are not affiliated with Recipe Blog Case Study, and figures are estimates, not verified data.
The strategy
1. Noticing citations and traffic had split apart
The site's recipes kept showing up in AI Overviews and ChatGPT answers for exactly the queries they targeted - rankings and citation appearances looked healthy in every rank tracker. But session counts kept drifting down quarter over quarter. The mistake had been treating "still cited" and "still visited" as the same metric. They were not: a full ingredient list and method is easy for an AI answer to quote in full, so a reader gets everything they need without ever clicking through.
2. Keeping the answer-first structure, but treating it as table stakes
They did not strip out the clear ingredient list or the direct method steps - those are still what earns the citation in the first place, and removing them would have meant losing visibility entirely. The shift was in expectation: answer-first content was reclassified internally as "the price of being found," not as a traffic strategy on its own.
3. Adding a tool that could not be summarized away
They added an interactive serving-size scaler and a unit converter (metric/imperial) directly on the recipe page - the kind of thing a text answer cannot reproduce, because it depends on the reader's own input. A short 40-second video showing the one genuinely fiddly step (folding, kneading, a visual doneness cue) was embedded above the fold for the same reason: AI can describe a technique in words, but it cannot hand someone the visual confirmation a video gives.
4. Splitting citation tracking from traffic tracking
They stopped watching "appears in AI Overviews" as their main success number and started tracking direct sessions and time-on-page per recipe as a separate line. The two moved independently: citation frequency stayed roughly flat month over month, while sessions on pages with the scaler tool and video climbed noticeably faster than pages without either.
Content structure
Answer-first recipe structure (kept, for citation) + interactive scaling/conversion tool + short embedded technique video + citation and traffic tracked as two separate KPIs
Results
AI citation frequency stayed roughly flat - the site was already being cited as much as it was going to be for these queries. Sessions on pages with the added tool and video rose meaningfully faster than on pages that only had text, without any change in how often those pages were cited.
Key lesson
A citation is not a visit. If an AI answer can fully satisfy the question with your text, most people will never click through, no matter how often you are cited. Keep the answer-first structure - it is still what earns the citation - but put at least one thing on the page that only delivers its value to someone who actually shows up: a tool, a calculator, a video, a downloadable asset. Then track citations and clicks as two different numbers, not one.
More case studies