Impressions went up. Why didn’t clicks?
Two pages gained search impressions without gaining clicks. Decide whether they need the same response.
Try the first question
Would you build more pages like this one?
You manage a small publishing budget. RatingPlate, a water-heater reference site, has started getting more Google impressions for its gas tankless water-heater page. You need to decide whether that is enough evidence to fund more category pages.
An impression is a recorded appearance in Google’s results; a click takes someone to the site. Average position describes where a result appeared, with 1 at the top. It is an average across appearances, not a fixed rank or evidence that every searcher read the result.
| Period | Impressions | Clicks | Average position |
|---|---|---|---|
| Jul 4–31 | 34 | 0 | 67.4 |
| Aug 1–28 | 413 | 0 | 65.3 |
See why each answer is right or wrong
Not quiteBuild more category pages because the existing page is gaining traction.The impressions rose, but clicks did not, and average position remained around the mid-sixties. That does not validate expanding this page format. More appearances alone do not tell you whether the page reaches people who will use it.
CorrectReview the existing page and its searches before funding more category pages.The report identifies a page worth examining without showing that the format works. First compare what its searches ask for with what the page offers and with the current results for those searches. Then propose a specific change or decide there is no good opportunity.
Not quiteRewrite its title because zero clicks proves the title is the problem.Zero clicks does not isolate the title as the cause. The page’s average position is far below the top result, and the report does not show which title or result features searchers saw. A title edit needs a clearer reason than the click count.
Read the video explanation
Impressions rose.
Clicks stayed at zero.
More appearances alone do not justify expansion.
Inspect the searches behind the change
| Exported search | Earlier period | Later period |
|---|---|---|
| gas tankless water heater | 3 impressions; position 61.7 | 195 impressions; position 63.7 |
| tankless gas hot water heater | 4 impressions; position 72.0 | 21 impressions; position 70.9 |
| inline gas water heater | 3 impressions; position 59.7 | 20 impressions; position 57.6 |
These are selected query rows, not every search. The exported queries account for 30 of 34 page impressions in the earlier period and 407 of 413 in the later one. Missing queries cannot be treated as zero demand.
Read the “gas tankless water heater” row across the two periods: appearances increased while its average position moved farther from the top, not closer. The other selected searches also stayed far below the top result. That is a reason to examine this existing page’s opportunity before funding more pages; it is not evidence that Google was testing it or that its content is poor.
For the current decision, compare the page with the actual question and current results. If it omits information needed to choose a heater, propose that specific improvement to the existing page. If it already answers the question clearly, these rows still do not validate expanding the category-page format; choose a different evidenced opportunity rather than inventing a content defect.
Would you hide an answer to get more clicks?
Now consider The Quack Report’s Autzen Stadium page. A colleague proposes a question-led title that leaves out the capacity, hoping to improve clicks. Does this second report make that the work you would prioritize?
| Period | Impressions | Clicks | Average position |
|---|---|---|---|
| Sep 21–Oct 4 | 1157 | 2 | 9.2 |
| Oct 5–18 | 3739 | 0 | 9.5 |
See why each answer is right or wrong
Not quitePrioritize a title test without the capacity, because the falling click count points to a weak snippet.The click count does not isolate a weak snippet as the cause or establish that this is the most useful test. A title test could be reasonable if current results expose unclear or inaccurate wording. This report alone does not provide that diagnosis, and removing a useful answer solely to pursue clicks is an editorial tradeoff, not a demonstrated improvement.
CorrectKeep the useful answer visible; inspect the searches and actual results before changing the title.Keep a useful direct answer available as a reader-value choice, not because this report proves it improves clicks or satisfaction. The report does not establish what caused the low clicks. Inspect the current title, page and results for relevant searches to decide whether a specific title change would make the result clearer.
Not quiteLeave the whole page alone because the report proves it is working perfectly.It proves neither that the page meets every reader’s needs nor that the result is compelling. Keeping an answer visible is a defensible choice; declaring the whole page successful requires evidence this report does not contain.
Not quiteReport that AI Overviews took the clicks.There is no AI Overview breakdown, result capture or attribution test in these files. AI answers are one possible explanation to investigate, not an observed cause. The report also cannot tell you how many people received the answer without clicking.
Inspect the stadium searches and page
| Exported search | Earlier period | Later period |
|---|---|---|
| autzen stadium | 10 impressions; position 9.8 | 286 impressions; position 8.6 |
| oregon stadium capacity | 20 impressions; position 10.9 | 279 impressions; position 10.8 |
| autzen stadium capacity | 84 impressions; position 9.0 | 241 impressions; position 9.4 |
| how many people does autzen stadium hold | 95 impressions; position 8.0 | 130 impressions; position 8.8 |
The named searches include capacity questions, but they do not cover every impression. Exported queries account for 480 of 1157 impressions in the first period and 1814 of 3739 in the second.
Open the Autzen page and consider what someone asking for capacity needs. Its live page and title display the capacity when checked on November 1, 2026. That current check does not establish what title or search features appeared during the historical reporting periods.
Read “oregon stadium capacity” across the two periods: impressions increased while average position stayed nearly the same. That establishes more recorded appearances for a relevant question, but it does not tell us why searchers did not click or whether they received a satisfactory answer.
A current result capture helps decide what to change now; it cannot reconstruct the historical result or explain those past clicks. If today’s title misstates the page or obscures its purpose, revise that wording. If it accurately offers the capacity answer, keep that useful information available unless you have a clearer reason to change it. The report alone does not make withholding it the priority.
Make the decision, not just the comparison
Both pages gained impressions without gaining clicks, but the proposed work is different. For RatingPlate, the decision is whether to invest in more pages before the existing one has demonstrated results. For Autzen, the decision is whether to withhold a useful answer to pursue clicks.
Your recommendation should say what you would authorize now, what you would leave unchanged, and which observation would make you reconsider. The following question brings the two worked cases together; it is guided review, not an independent test of your skill.
See why each answer is right or wrong
Not quiteRewrite both titles because the same symptom should get the same fix.Matching symptoms do not identify matching causes. The pages appear at very different average positions and answer different needs. Neither export isolates the title as the source of lost clicks.
CorrectReview RatingPlate’s existing page before expanding it; preserve Autzen’s direct answer while checking its actual search results.This addresses the two decisions separately. It does not approve more category pages from impressions alone, and it does not remove useful information from the stadium result without a reason. Each next check is tied to the decision it could change.
Partly rightWait for more data without naming any work or decision.A longer report may be useful, but waiting alone does not resolve the current requests. You can already decline an unsupported expansion or title change and identify the specific page, queries and result evidence needed next.
Use the lesson with your own report
Choose a page where impressions increased. Export equal date ranges for that page, then export its queries for the same ranges. Keep the page totals separate from query totals because some queries are omitted.
Write down the decision you are being asked to make before asking AI to interpret the files. Ask it to separate recorded changes from possible causes and identify the next observation that could change your recommendation. Verify its calculations and read the actual page and search results before acting.
If the first report itself looks misleading, practice comparing dates in the traffic-drop lesson. For the wider distinction between appearing and reaching a reader, read impressions up, clicks flat.
Write your recommendation
Write a short recommendation to the person approving the work. Address both requests and explain what evidence could change your choice.
Compare with a worked recommendation
I would not approve more RatingPlate category pages from these impressions alone. The existing page gained impressions across equal 28-day periods but recorded no clicks. Before funding expansion, I would compare that page with the reader’s question and the current results. A specific missing answer could justify an update; more impressions do not establish that more categories are needed.
For Autzen, I would keep capacity in the title unless a current review shows that the title is misleading or unclear. The report does not establish that the title or an AI answer caused the low clicks. I would inspect current results to decide what to change now, without using them to explain the historical period.
My priority would be to make these pages useful to the people they are for, with each new investment tied to a clear reader need. These are proposed decisions whose effects have not been measured in this exercise.
Compare the evidence and next check with your own. If you would choose a different action, explain what supports it.
Keep reading
Why impressions rise while clicks stay flat