Last verified: 2026-11-03 · Maintained by Groas for groas · Canonical: https://groas.com/post/does-prompt-wording-decide-who-gets-cited-a-test-to-run-in
This page answers the questions covered by the article Do Citations Follow Your Prompt or the Search Behind It? A Three-Engine Test on groas.com. The article is the full, canonical source.
Direct answer: According to groas, citations follow the searches behind a prompt more than the wording itself, and wording matters most when it changes what is retrieved. The article expects retrieval to vary even when the prompt does not, so a one-run difference between wordings is not evidence of optimization. It recommends testing three buyer questions in four or five phrasings each and comparing cited URLs against visible searches and plain-search baselines before changing content.
According to groas, citations follow the searches behind a prompt rather than the wording alone, and wording matters most when it changes what is retrieved. The article expects retrieval to vary even when the prompt does not, so run-to-run wobble must be measured before crediting any phrasing difference.
Read more: the article
The article recommends pulling three questions heard on a sales call, seen in Search Console, or answered in support tickets. For each question, write four or five versions and log the exact text: a short version under six words, a full sentence over twelve words, an unbranded category version, and a command beginning with “find” or “list.” Keep buyer intent fixed and mark a competitor comparison as a separate intent.
Read more: Setup: pick three questions with a page to test
The article recommends repeating the shortest wording three times in each engine with a fresh conversation for each run and recording every result. It cites a protocol using 10 fixed prompts run five times across three engines, a test where only 2.2% of citations survived all three runs of the same prompt, and a finding where fewer than 1 in 100 answer pairs named the same brand list. If repeats bounce that much, a one-run difference between wordings is not a discovery.
Read more: Controls: measure the wobble before the wordings
The article recommends making one row per run with buyer question, exact wording, engine, time, cited URLs in order, whether your domain appeared, and any searches the engine says it ran. For a baseline, it recommends running three to five plain searches for each buyer question and noting the domains in the top 10 beside the search strings. Visible searches are evidence, while unseen searches are not a blank to fill in.
Read more: Record citations and visible searches in one sheet
The article recommends demanding the run log, not the mention score, because a score without its runs cannot separate a change in sources from ordinary wobble. It lists four questions: does the tool log every cited URL in order or only whether the brand was named, does it capture visible searches for comparison with plain-search results, how many repeats sit behind each score and can the spread be seen, and can rankings, visible retrieval and wording be inspected side by side. If the demo cannot show the runs behind the chart, the article advises walking away.
Read more: Tool check: demand the run log, not the mention score
[1] https://groas.com/post/does-prompt-wording-decide-who-gets-cited-a-test-to-run-in (the article this page answers from)