How to Check If a Keyword Is Worth Writing About (Before You Spend a Day on It)

By Katleen, Founder of PARCHMNT · Updated October 29, 2026

To check if a keyword is worth writing about, read Google's first page for it before you write a word: who holds the ten spots, what kind of pages they are, and whether an AI answer sits on top. A newer site can win when the search is specific and some spots are held by forums, videos or thin pages. It cannot win a one-word topic held by encyclopedias and giants.

The reading takes about ten minutes and needs no paid tool. It is keyword research without tools, and it is more honest than the dashboard version. This post explains why the tool numbers are weaker than they look, what to look for on the page, and the four verdicts the reading gives you.

Why can't you trust search volume and difficulty scores?

Search volume and difficulty scores are estimates built from models, not measurements of your chances, and the tool makers say so themselves. The numbers summarise links and yearly averages; they do not look at who holds the spots today.

Ahrefs describes its search volumes as "still just estimations rather than exact values" and notes that most tools report an annual average, so a December spike and a July lull become one flat number. On difficulty, Ahrefs' Tim Soulo writes that the score "is merely an estimation" and that it "doesn't account for your specific website's authority," because it is built mostly from the backlinks of the pages that rank.

Semrush's own knowledge base describes its KD% as a prediction of how hard it would be to rank in the top ten, built from referring domains, link ratios and authority scores, plus qualities of the results page itself.

What that means for you: a "difficulty 20" keyword can be hopeless if the ten spots belong to government pages and the biggest publisher in your field, and a "difficulty 60" keyword can be open if half the first page is forum threads. The number summarises links; it does not look at who is actually there, and who is there is the thing that decides.

What does the first page tell you?

The first page tells you three things the numbers can't: who holds the spots, what kind of pages they are, and what the searcher is likely to click. Write them down in a plain table.

• Who holds the spots. Count how many of the ten results belong to encyclopedias, major publishers, big retailers, platforms and official (.gov or .edu) sites. A newer site will not outrank those head on. Then count the openings: forum threads, Q&A sites, videos, and thin or old blog posts. A complete, well-sourced page can pass those.
• What kind of pages they are. If the first page is all product listings, Google has decided the search is about buying, and an article will not fit. If it is all long guides, a short answer will not either.
• What is above the results. An AI Overview or a featured snippet takes clicks even from the top spot. First Page Sage's 2026 study of 1,184 domains puts the click rate for position one at 22.6 percent when no AI answer is present and 3.6 percent when one is. That six-fold gap means a search with an AI answer is worth writing for only if your page can be the one the answer cites, which is a different page from one written only to rank.

How specific should the keyword be?

Specific enough that you can tell what the person wants. One word ("composting") is a topic, not a search anyone wins, because the ten spots go to encyclopedias and the searcher's goal is unknown. Two words with no question in them ("time tracking") are usually the same. A phrase with a question, a comparison, a cost, a place or a "for" is a search: "how to track billable hours as a freelancer", "compost tumbler vs bin", "family dentist in Aurora".

One search per page. A page written for three keywords usually wins none of them, because Google ranks a page for the search it answers best, and assistants cite the page that answers one question completely. The related phrases come along on their own when the page is thorough; they do not need their own page.

What are the four verdicts?

The reading ends in one of four verdicts: too broad, saturated, competitive or winnable. These are the rules PARCHMNT's keyword check uses, and they are simple enough to apply by hand. If a post is already live and not moving, why a blog post is not ranking walks the same checks from the other end.

PARCHMNT is a writing tool that researches what ranks for a search, writes the page, checks every claim and tracks where it lands in Google and AI answers. Its check reads the first page once, the way a person would, and the verdict is computed by code from those counts rather than asked of a model. We built it that way because the reading is the part that needs eyes, and the judgement is the part that needs rules. In our own reading for this journal, two of the ten long-tail searches we considered were held by vendor blogs and glossaries and went straight to the saturated pile; the five we wrote had forum threads in their top ten.

What should you write instead when a keyword is saturated?

Write the narrower questions the same page shows you: the "People also ask" box and the related searches at the bottom. Keep the ones that share your subject, are specific, and are not the keyword itself. "Composting" becomes "why does my compost smell" and "compost tumbler vs bin"; "time tracking" becomes "how to round billable hours".

Narrower searches like these are usually where forum threads rank, which is the clearest sign of all that nobody has written the complete answer yet. A cluster of five narrow pages, linked together, also builds the authority that lets you come back for the broader search later.

How do you estimate the payoff without search volume?

Estimate the payoff from position, not from volume: the click rate at the position you can realistically reach, times a rough sense of how many people ask. A winnable keyword at position three is worth more than a saturated one you would reach at position fifteen.

First Page Sage's 2026 figures give a blended click rate of 7.1 percent at position one, 1.7 percent at three, 0.7 percent at five and 0.2 percent at ten, across all searches, with and without AI answers. The lesson in those numbers is that page one is where clicks begin, and the top three is where most of them are.

The second half of the payoff is being cited. Ahrefs found in July 2025 that 76 percent of pages cited in AI Overviews ranked in the top ten, and its March 2026 update put the share at about 38 percent as Google widened where it draws from; either way, a top-ten page is the one most likely to be cited. Pew Research Center found in 2025 that people clicked a traditional link in 8 percent of visits with an AI summary against 15 percent without. So a page that reaches the top ten and gets cited keeps its clicks; a page that reaches the top ten and is not cited loses about half of them to the summary. The keyword is worth targeting when both are possible, which is why the page has to be written to be quoted by AI, not only to rank.

Key takeaways

• Tool scores summarise links and averages; the first page shows who is actually there.
• Count the spots held by major sites and the openings held by forums, videos and thin pages.
• One or two words with no question in them is a topic; a question, comparison, cost or place is a search.
• Four verdicts: too broad, saturated, competitive, winnable. Write the winnable ones and the narrower searches under the saturated ones (short tail versus long tail for a small business shows what each looks like).
• An AI answer on the page cuts the top spot's clicks from about 23 percent to under 4 percent, so aim to be the cited source, not only the ranked one.

Frequently asked questions

Is a paid keyword tool worth it?

For a small site, usually not at first. The tools are good at listing phrases you had not thought of, and weak at telling you whether you can win them, by their own account. Read the first page for the ten phrases that matter most to your business before paying for data on a thousand that don't.

How many keywords should one page target?

One. A page wins the search it answers best. Related phrases and questions belong in the body, where a thorough page covers them naturally, and a second keyword that deserves its own answer deserves its own page, linked to the first.

How long does it take to see whether a keyword worked?

Google Search Console shows a new page's position within a few days to two weeks of indexing, and the position usually settles over two to six months. Judge the keyword at four weeks (is it on page one or two?) and again at twelve.

What if the first page is all forum threads?

A first page of forum threads is winnable, and probably more valuable than it looks. Forum threads rank when nobody has written the complete answer. A page that answers the question properly, with sources, usually passes them within weeks.

How to check if a keyword is worth writing about, in ten minutes

Read the page, count the spots, pick the verdict, and write only the winnable ones. If you want the reading done for you, with the verdict and the narrower searches, try one article free; the check runs before the article is spent.