High Impressions, Low Clicks: Our Real Case Study

A page from Agência Primeira Página ranks third on Google and gets a 0.7% click-through rate. In 28 days, that was 856 impressions and 6 visits.

This isn't an isolated case for our site. It's what happens with any content that ranks well for a question Google already answers on its own — and it's a costly mistake, because it looks like success in every report.

We're opening up our own numbers because almost no one publishes theirs.

The numbers, unfiltered

28-day window, from August 8 to September 4, 2026, straight from Search Console:

• Entire site: 12,001 impressions, 130 clicks, 1.08% click-through rate, average position 10.
• A single topic — the meaning of the expression "career in decline" — accounted for 4,420 impressions and 13 clicks. That's 36.8% of everything the site appeared for that month.
• The main page for that topic: position 2.9 and 0.7% CTR.
• 93 typed variations of the same expression added up to 2,866 impressions and 2 clicks.
• Excluding that topic, the site is left with 7,581 impressions and 117 clicks: the click-through rate rises from 1.08% to 1.54%.

In other words: a third of our "audience" on Google doesn't exist as an audience. It inflates the chart and never walks through the door.

Why the high ranking didn't turn into a visit

Because the search was a dictionary-type query. Someone who types "what does 'career in decline' mean" wants a definition, and Google delivers it right on the results page — in a featured snippet, an AI summary, or the excerpt below the title. The person was already served before ever needing to click.

Notice the shape of the searches that generated these impressions: "career in decline," "decline of a career," "career decline meaning," "figuratively speaking, saying a career is in its decline means it is." These are vocabulary questions. None of them come from someone looking for a vendor.

And here's the uncomfortable part: the copy isn't bad, and the ranking is great. The problem happened before writing even started, in the choice of keyword.

The test that avoids this mistake

There are four checks, and they take ten minutes per keyword:

1. Type the keyword into Google and see who's already there. If the first page is a dictionary, encyclopedia, news portal, or government site, the search is informational. If it's competitors, catalogs, and service pages, it's commercial.
2. Check whether the answer appears before the results. A featured snippet at the top or an AI summary means the click has already been consumed.
3. Look at autocomplete. Type the keyword and watch what Google suggests. If the suggestions are "what is," "meaning," "synonym," the intent is to understand. If they're "price," "how much does it cost," "company," "near me," the intent is to hire.
4. Ask what the person's next action would be. After reading your copy, would they request a quote? If the honest answer is "no, they just wanted to know," the keyword isn't for you.

Two more cases of ours, measured

"Escola 3.0" (School 3.0). We had a service page targeting this term, and it does bring in real impressions. We checked the results page: the top spot belongs to the Rio de Janeiro City Hall, and the related questions are things like "how to log into 3.0." It's teachers looking for the login to the Department of Education's system, not principals looking for a vendor. We dropped the term from our strategy.

"Captação de alunos" (student enrollment). Same sector, opposite result: the results page is entirely companies selling services to schools. It's commercial, and we built a post around this term.

The lesson from both checks is the same: a synonym is confirmed on the results page, not in the dictionary. Two terms that seem like siblings can have completely different audiences.

What to do when it's already happened

Don't delete anything. Informational traffic doesn't pay the bills, but it doesn't cost anything either. Here's what you can do:

• Leave the copy live and stop investing in it. No rewrites, no new links, no team time.
• Link the page to something commercial with an honest internal link, for anyone who happens to be interested.
• Take the topic out of your performance reading. As long as it's mixed into the average, you'll misread the whole site — that's what happened with our click-through rate, which looked like it was falling when it was actually stable.
• Redirect your effort to the keywords the check approved.

What to measure from here on

Average position is a means, not a result. The three metrics that actually decide something:

1. Clicks per topic, not for the whole site. The average hides exactly the case described in this article.
2. Click-through rate per page. A high ranking with a low rate signals the wrong intent, not a bad title.
3. Leads that came in. It's the only number that pays salaries.

If you want the other side of this coin, we've already told the story of how we went from 20 to 161 clicks, and how to rank at the top of Google Maps remains the shortest path for businesses that serve a specific region. Choosing keywords with commercial intent is the work we do at agência de SEO.

And the article that generated these numbers is still live, exactly as it was. It answers a question well that no one ever needed to click to have answered.