I was reviewing content ideas for WhittleOS and found a phrase that looked impossible to ignore. The trend line was steep. The wording seemed adjacent to our product.
Then I read the result page.
The searcher was not a solo founder trying to decide what to build. They were looking for corporate problem-solving frameworks, consultancies and process methods. A WhittleOS article would have had to smuggle a startup topic into a different question.
So we killed it.
That changed how I read keyword data. Growth describes a query, not a customer. Before a topic enters the queue, I now write down who searched, what job they need done, and whether our page would be the best answer even if it never mentioned WhittleOS. If any part is fuzzy, volume does not rescue it.
The review and the gates we now use are here: https://whittleos.com/guides/search-intent-vs-search-volume
The searcher wasn’t wrong, though, they were just solving a different problem. Killing the article makes sense, but I wouldn’t discard the signal.
If several rejected queries cluster around the same adjacent job, that could point to a new customer segment or product opportunity rather than a content opportunity. Do you keep the rejected terms and reasons somewhere, then review those clusters later?