Case study: 239 clicks from 61,500 impressions — anatomy of a page-2 CTR collapse

The setup

An indie-SaaS site, ~18 months old, content-led acquisition, no paid spend. The owner opened Search Console expecting validation and found this panel instead. They sent me a screenshot with the message "is this as bad as I think it is?"

It is, and it isn't. Worth walking through because the diagnosis is the same one I give roughly two-thirds of the indie devs who hand me a GSC tab.

The data

Trailing 28 days, web search type:

Metric	Value
Total clicks	239
Total impressions	61,500
Average CTR	0.4%
Average position	11.5

And four things visible in the daily chart above:

• A spike on days 1–4 (~28 clicks/day at peak).
• An immediate crash to 2–3 clicks/day for about a week.
• A recovery on days 14–18 to ~20 clicks/day.
• A slow decline back to ~5 clicks/day across days 19–28.

The instinctive read is "I need to rank higher." That's half right, and it's mostly the wrong place to start.

Diagnosis, in three reads

Read #1: position 11.5 means you mostly live on page 2

Google still ranks results in groups of 10 per page. Average position 11.5 means the typical time someone sees this listing, it's the first or second result on page 2.

The gap between position 10 (last on page 1) and position 11 (first on page 2) is the largest single-position drop in the entire CTR curve. Aggregate studies consistently put position 10 around 2.5% CTR and position 11 around 1.5% — and that's already generous, because in 2026 most users genuinely never scroll past page 1.

So 61,500 page-2 impressions are functionally invisible. They're "displayed" the same way a billboard behind a building is displayed.

Read #2: even for page 2, the CTR is still ~3x too low

The red dashed line on the chart is the math. At average position 11.5, the expected CTR for an unremarkable listing with no SERP-feature interference is roughly 1.2%. That works out to about 26 clicks per day at this impression volume.

The actual daily clicks line spends almost the entire 28 days under the red reference line. Most days it's at less than a quarter of expected.

• Expected at 1.2% CTR: 61,500 × 0.012 ≈ 740 clicks
• Actual: 239 clicks
• Gap: ~500 clicks left on the table from listing copy alone, not ranking.

Translation: even on the days the listing surfaces above the fold, people are looking at it and choosing someone else. That isn't a ranking problem. That's the listing itself.

Three common causes for listing-level CTR collapse:

1. The title doesn't echo the query. When a user searches "best AI coding agent 2026" and your title is "A Guide to Autonomous Developer Tools," their eye doesn't pattern-match. They scroll past you. Title rewrites that echo the user's actual phrasing routinely 2–3x CTR with zero ranking change.

2. The description is missing or auto-generated. If you don't ship a deliberate <meta name="description">, Google synthesizes one from your H1 + first paragraph. The synthesized version is almost always less compelling than what you would write on purpose. Google still rewrites your description ~70% of the time, but the 30% it leaves alone is the highest-CTR third of your inventory.

3. SERP features are eating the click before it reaches you. AI Overviews, featured snippets, People also ask, shopping carousels, video thumbnails — all push organic listings down and consume clicks above you. At position 11.5 in a SERP with three rich features, you are effectively at position 18 from the user's perspective. No amount of title polish recovers that; the page has to target a different query.

Read #3: the shape tells you when, not why

The four-act pattern annotated on the chart — spike, crash, recovery, decline — is a fingerprint. It almost always points to one of three things:

• A single high-volume query you ranked for dropped out. The most common cause in 2026: an AI Overview now answers the query, so even position 1 gets a fraction of the clicks it used to.
• You briefly owned a featured snippet and lost it. Snippet wins look like step-changes up; losses look like cliffs down. The crash on day 5 here is consistent with a snippet loss.
• An event or news cycle drove temporary demand that has now faded. Common for opinion pieces, X just launched coverage, anything seasonal.

To figure out which: in Search Console, hit Compare → last 14 days vs. previous 14 days, sort by impression delta descending, look at the top 5 dropped queries. That ranked list is the story. Whatever query dropped most is the answer.

The full diagnosis, in one paragraph

This site has demand. 61,500 impressions in 28 days is not a discoverability problem — Google thinks it's relevant for a lot of searches. What it has is two stacked execution failures: rankings sit just outside the page-1 cliff where clicks would be 5–10x higher, AND the listings that do show up are converting at roughly one-third the rate they should at that position. There's also a specific event in the middle of the window worth diagnosing, but it's a secondary thread.

Action plan: what I'd do tomorrow morning

In order, because the order matters:

1. Filter GSC for impressions > 1,000, CTR < 1%, position 8–15. This is the highest-leverage list in the entire account. Every row is a page that's a single rewrite — title + meta description — away from a 3–10x click multiplier without any new content or new links.

2. Rewrite titles to echo the actual top query for each page. Pull the top 3 queries for the page in GSC, find the shared intent, rewrite the title in the user's own language. Stop being clever. Be specific.

3. Write deliberate meta descriptions for the top 10 high-impression pages. Treat them like ad copy, not summaries. The 30% Google leaves alone is your highest-CTR third.

4. For the position 8–15 cluster, add one internal link each. Pick your strongest existing page (by impressions in GSC) and add a contextual link to each underperformer. Internal linking is the single fastest ranking lever you fully control.

5. Check the SERP for your top 5 queries by hand. Actually open google.com, search the query, screenshot the SERP. Count the rich features above the first organic result. If there are three or more, you will never get more than ~0.5% CTR from that query no matter what you do — pivot the page to a different query instead of trying to win this one.

6. Diagnose the mid-period drop. Compare mode in GSC, last 14 vs. prior 14, sorted by impression delta. Find the query. Decide if it's recoverable, or if it's now permanently owned by an AI Overview.

The lesson (it's a generalizable one)

Most "my SEO is broken" problems indie devs bring me are CTR problems wearing a ranking-problem costume. The reason is structural: every analytics tool reports ranking as the headline number because it's the easiest to measure, so we end up obsessing over a metric that explains a much smaller share of the click outcome than we think.

Rankings decide if you show up. Title, description, and SERP context decide if you get clicked. Both matter. In most underperforming accounts the second is the one with 5x more leftover money on the table.

If your Search Console looks anything like the chart above — big impressions, tiny clicks, position in the teens — don't go write more content. Don't go buy links. Open the high-impression / low-CTR view first. The fix is almost always smaller, cheaper, and faster than you think.