One Page Drove 62% of Our Google Clicks: A Case Study

Direct answer

In this one-site case, from August 29 to September 25, 2026, one Perplexity review received 243 of 390 Google Search clicks to konabayev.com: 62.3%. These are Google clicks to a review about Perplexity, not visits from the Perplexity assistant. The site total rose slightly versus the previous equal 28-day window, but clicks to every other page combined fell 196 → 147 (25%). This is a secondary analysis of the same GSC export used in our AI visibility case study, not an independent market benchmark or a safe-concentration threshold.

The page’s absolute gain was larger than the rest-of-site loss. “Clicks increased” and “most of the portfolio weakened” are both true. This is why a founder should receive the page decomposition along with the site total.

A simple sensitivity calculation

Imagine that the current leading page loses half its clicks while the rest stays at 147. The hypothetical total is 121.5 + 147 = 268.5 clicks (about 269), 31.2% below 390. This is a stress-test scenario, not a forecast. The 50% shock is chosen for illustration, not estimated from a model; other pages could also change.

Use the downloadable concentration worksheet to reproduce the two windows and vary the scenario. The fields are period, top_page_clicks, other_page_clicks, total_clicks, top_page_share and hypothetical_total_if_top_page_halved. A portfolio with more than one strong landing page is less sensitive to one URL, but a page count by itself does not prove resilience. You need page-level click or conversion shares.

What the pattern tells us to investigate

First, separate demand and ranking. For the affected pages, export clicks, impressions, query mix and average position in equal complete windows. A click change with roughly stable position calls for a different investigation from one with a large position change; neither pattern alone proves a cause.

Second, compare queries and landing pages. A site can gain clicks because one product name surges while losing the nonbranded, commercially relevant questions that feed leads. Preserve the same country, search type and calendar-day boundaries across both exports. Group URLs by their real editorial intent: review, comparison, how-to, data, service page. Do not call a click a customer or add GA4 sessions to GSC clicks.

Third, inspect conversion outcomes. Our recent PostHog snapshot contained two main CTA clicks, one affiliate click and zero email signups in seven days. Those counts are too small to establish revenue direction; they do show why the next reporting layer needs a defined action and a longer window.

What we would publish each month

Report total GSC clicks, top-one and top-three page shares, clicks outside those pages, and qualified actions from the corresponding landing pages. For a change, list the pages and queries that moved. A top-page share is a risk indicator only in context: a focused product site may reasonably have a dominant page, while a broad editorial site could be exposed if its only winner fades.

For measured assistant referrals, use our GA4 and PostHog workflow. Our five-metric report keeps click concentration beside separate visit and outcome rows. Cite this result as “Konabayev.com, GSC Web clicks, August 29–September 25, 2026, 243 of 390 clicks to one review (62.3%); one site, not a market benchmark.” The CSV contains the two dated windows and hypothetical calculation.

What this dataset cannot establish

Concentration describes the click distribution; it does not identify a cause or prescribe a safe threshold. The export contains two equal windows from one website. It does not show a representative sample of publishers, a counterfactual for a site without the leading review, or which future topic will earn equivalent demand. The 50% top-page reduction in the CSV is intentionally hypothetical. An editor who cites it should label the scenario and the observed rows separately. Google’s Search Console performance documentation explains the Web click unit used here.

For a decision, break the −49 clicks outside the leading page into affected URLs and query clusters. Then ask whether the lost clicks were branded, informational or buyer-intent and whether their associated qualified actions moved. This prevents a site-total increase of ten clicks from concealing a decline in pages that matter more to revenue. It also prevents a 62.3% top-page share from being treated as a universal danger line. A focused site can have a rational dominant page; the risk is an unexamined dependency on that page’s continued performance.

FAQ: click concentration

Is a 62.3% share good or bad?

There is no universal answer. It means one page received 243 of 390 Google Web clicks in the specified 28-day window. For a site built around one product or topic, a high share may match its purpose. For a broad editorial portfolio, it can make the total vulnerable to one query cluster. Judge it with the page’s commercial value, other pages’ trajectories and qualified actions. The two-window case is an illustration, not a cross-site benchmark.

Are these clicks referrals from the Perplexity assistant?

No. The leading URL is a review about Perplexity, but the 243 clicks are from Google Search results reported in GSC. A visit referred by perplexity.ai is a different event in an analytics source report. Mixing them would make the review’s Google demand appear to be assistant referral growth. The Perplexity signal guide separates those definitions and the relevant measured windows.

Why calculate clicks outside the top page?

The site total rose 380 → 390, which looks stable. Removing the review’s 184 → 243 reveals the other pages moved 196 → 147. This decomposition is arithmetic, not a causal model. It tells the team where to investigate next: pages and queries outside the winner. It cannot show whether the change came from ranking, demand, snippets, seasonality or measurement without further page and query evidence.

Last verified: September 2026. The CSV is a dated single-site Google Web click analysis.