Search volume tells you how often a phrase was estimated to appear. Search intent tells you who is behind it and what they need next. For a product site, intent wins. In our own keyword review, one fast-moving phrase was dropped, one enormous-looking term became only an experiment, and one modest commercial query kept its page because conversion—not traffic—was the job.
A keyword plan is not a publishing queue
We received a founder-supplied Keyword Planner workbook and used it to reorder the WhittleOS content backlog. We did not use it to authorize pages. That distinction sounds procedural until a sorted sheet puts the wrong work at the top.
Google describes average monthly searches as an estimate based on the selected date range, location, and Search Network settings, averaged over the period—by default, a year. That makes it useful for comparison and dangerous as a direct observation. The number does not tell you whether one month carried the average, who made the search, or what kind of page satisfied them. Google's own definition leaves those questions to you.
Our workbook figures stay out of this page. We could verify its sheets and formulas; we could not verify the original exports, settings, or merge. A modeled planning input can decide what to inspect first. It cannot become a public claim merely because the cell is precise.
The four checks that can reverse the ranking
Volume starts the review; none of these checks can be answered by volume. A failure at any one changes the page, its measurement contract, or the decision to publish at all.
Language comes first. A phrase can contain every word in your positioning and still belong to a different profession. Search engines resolve jobs, not your private definition of the words.
Then inspect the distribution. Averages compress time. If one exceptional month dominates, the headline describes arithmetic correctly and demand badly. Current months, the median, and the shape of the series tell you whether the summary is stable enough to act on.
Read the result page by hand. The ranking pages are the engine's current answer to who this searcher is. If they are management consultancies and process guides, your startup-idea page is not an ingenious alternative. It is an irrelevant result.
Finally, name the asset's job. Acquisition pages earn reach. Conversion pages help a smaller set of qualified visitors decide. Proof pages make a claim checkable. Judging all three by impressions rewards broad pages and deletes the pages closest to revenue.
Three decisions from the same sheet
| Phrase | What looked attractive | What inspection found | Decision |
|---|---|---|---|
| company problem solving | Recent movement put it near the front of the planning queue. | The result page was corporate problem-solving methodology: consultancies, business schools, Six Sigma, and workplace process guides. | DROP |
| small entrepreneur ideas | The headline average made it look like the largest acquisition opportunity. | The month-by-month series showed that one exceptional month dominated the average, so the summary cell was not a useful picture of current demand. | EXPERIMENT |
| startup idea validator | The query looked modest beside broad idea terms. | The searcher is comparing a product category and is close to choosing a tool, so the page has a conversion job rather than a traffic job. | KEEP |
Review completed 2026-08-19. The figures that prompted it remain in the planning document; the decisions and the inspection that justified them are what can travel.
The phrase we dropped because the searcher was wrong
“Company problem solving” looked adjacent to our thesis: find a business problem, then decide whether it supports a product. The result page meant something else. It was filled with corporate problem-solving methods, business-school frameworks, Six Sigma, and consultancy process guides. The searcher was a manager trying to solve a problem inside a company, not a founder looking for a problem to build around.
There is no copy adjustment that fixes that mismatch. Publishing would buy the wrong visit even if the page ranked. We dropped it.
The average we demoted to an experiment
“Small entrepreneur ideas” sat at the other extreme: broad language and a headline average large enough to pull the plan toward it. The monthly series told a different story. One exceptional month dominated the summary, while the phrase itself reached a much broader reader than the solo technical founder WhittleOS serves.
That does not make the term fake. It changes the size of the bet. A single landing-page experiment can test whether this broader audience moves toward our method. Rewriting the product, navigation, and editorial calendar around it cannot be justified by a volatile planning estimate.
The modest query we kept for a different job
“Startup idea validator” looked small beside the idea-list terms and stayed important because it is commercial. The reader is not browsing for inspiration. They are deciding which tool to trust.
That is why /startup-idea-validator is a conversion page rather than a traffic pillar. It carries the method, the known misses, and a real report. If it attracts fewer but more qualified visits, it can succeed without ever winning the acquisition contest. Killing it for low reach would be measuring a checkout by footfall.
Match the measurement to the page
Judge qualified discovery: impressions, clicks, and whether the searcher belongs to the ICP.
Judge the next action: signup, a completed first run, or movement toward checkout.
Judge whether the reader can inspect the receipt and understand the limitation without us.
The same keyword metric cannot judge all three assets. Name the job before choosing the success measure, or the broadest page wins every internal argument.
This is also why publishing a separate page for every keyword variation is the wrong response. Google recommends people-first content and explicitly warns against producing many pages across topics merely to attract search visits. Its guidance also says to write for an existing or intended audience and to contribute original experience rather than a summary of what others already wrote. That is the useful bar here, not whether a spreadsheet has another row. Read the guidance in full.
The operating rule
Use keyword data to decide the order of inspection. Then make the publishing decision from four other things: meaning, monthly shape, the actual result page, and the job the asset performs for the business.
If the searcher is wrong, drop the page. If the distribution is unstable, shrink the bet to an experiment. If the query is small but commercial, keep the page and judge conversion. The sheet is an input to the decision, not the decision wearing decimals.
The short version
- Keyword Planner provides estimates shaped by date range, location, and network settings; it does not identify your buyer or page strategy.
- Our 3 cases produced 3 different outcomes from one planning sheet.
- A fast-moving term died on intent. A huge-looking average became an experiment. A modest query stayed because it has a conversion job.
- Search figures order the review. The result page and the product decide what ships.