Fourteen Queries Took 93 Per Cent of the Labelled Clicks. The Organic Growth Strategy Runs Backwards.

TLDR: A year of Google Search Console data across three Swiss properties Kainjoo manages shows fourteen branded queries taking 93 per cent of every click a query label can attribute, which puts the four-stage organic growth strategy in the wrong order.

Fourteen queries took 93 per cent of the labelled clicks

Between 1 August 2025 and 31 July 2026, one Swiss business-to-business (B2B) property inside Kainjoo’s own group, the only one of the three left unnamed here, logged 144,056 impressions in Google Search and 1,943 clicks at an average position of 13.4. A branded vs non-branded search split is the first cut worth making in that data, and it removes the floor.

Search Console named 555 queries across the year, covering 18,218 impressions and 891 clicks. Twelve of those queries spell the company name or a product name correctly. A branded query carries the company or product name; a non-brand query does not, and that single line is the whole split. Those twelve produced 4,343 impressions and 815 clicks, a click-through rate (CTR) of 18.8 per cent. The remaining 543 queries, every one a phrase a content programme is built to win, produced 13,875 impressions and 76 clicks.

So 91 per cent of the clicks Search Console can attribute to a query came from people who already knew the name. One limit governs that number. Search Console drops rare queries from the query dimension to protect user privacy, so those 555 named queries carry 18,218 of the property’s 144,056 impressions and 891 of its 1,943 clicks. That concentration sits inside the 46 per cent of clicks carrying a query label. Against every click the property earned, the twelve branded queries took 42 per cent. The Search Analytics application programming interface withholds the same rows the Performance report does, and the totals shown by search type are the only complete figures on the property.

One correction sharpens it further. Two queries sitting in the non-brand bucket, “kainjo” at 54 impressions and “kanjoo” at 504, are misspellings of the company name, and a misspelling of a company name is a branded query wearing a disguise. Reclassifying both moves the branded set to fourteen queries, 4,901 impressions and 833 clicks at a CTR of 17.0 per cent, or 93.5 per cent of every labelled click. The non-brand set drops to 541 queries, 13,317 impressions and 58 clicks, a CTR of 0.44 per cent, with 504 of those queries producing no click at all. The error ran against the argument, and fixing it made the concentration worse.

The received model says awareness first, then capture, then conversion, then expansion and advocacy. On this property the arithmetic runs the other way. The last stage, a name the buyer already holds, is the only one producing measurable clicks. The three in front of it produced 58.

The sequence broke in 2009, and McKinsey said so

Stage models inherit their shape from the advertising funnel, under attack from inside consulting for seventeen years. McKinsey Quarterly published the consumer decision journey in June 2009 and argued flatly that the funnel fails to capture touch points that decide a purchase. Two findings from that work land directly on the data above. First, brands present in a buyer’s initial consideration set are up to three times more likely to be purchased than brands that are absent. Second, during active evaluation, two-thirds of touch points are consumer-driven rather than company-driven.

A branded query in a Search Console export is the initial consideration set showing up in a log file, the buyer typing the name they arrived with, misspellings included. Everything a stage model files under “awareness” is an attempt to get into that set, and the click curve says the attempt fails at the click, long after the ranking was won.

Swiss firms feel this harder than most. The Federal Statistical Office counts more than 600,000 enterprises in Switzerland, of which 99.7 per cent are small and medium-sized enterprises (SMEs) and roughly 90 per cent employ fewer than ten people. The State Secretariat for Economic Affairs adds that SMEs are more than 99 per cent of commercial enterprises and two-thirds of jobs. A market of micro-firms generates very few searches per topic, which is the second half of the problem.

The non-brand half of the branded vs non-branded search split earned 58 clicks

Bucket the corrected non-brand queries by average position and the click curve is brutal. Between positions three and six, 1,841 impressions produced 18 clicks. Between six and eleven, 5,110 produced 25. In the top three positions, where the industry assumes the money sits, 69 impressions produced none.

Two more caveats. The bands mix intents, and a query at average position four here is usually a hard commercial phrase, so position explains only part of the collapse. And Search Console gives no way to isolate it, because Google states that artificial intelligence (AI) features are included in the overall search traffic and reported inside the “Web” search type of Search Console. AI Overviews impressions and AI Mode impressions are already inside those 13,317. One limit outranks those three. This is a multi-brand group domain ranking for a portfolio of its own names, so its branded share sits above what a single-product SME should expect. All three share one content programme, so part of that collapse may be ours. The AI tax is being paid; the receipt does not itemise it.

Pew measured yield per search, Google measured aggregate volume

The public argument about that tax has two evidenced sides. Pew Research Center instrumented the browsers of 900 United States adults and analysed 68,879 Google searches during March 2025. Pew found users who met an AI summary clicked a traditional result in 8 per cent of visits, against 15 per cent when no summary appeared. Clicks on links inside the summary itself happened in 1 per cent of visits. Some 18 per cent of the searches in the sample generated a summary at all.

Google’s answer, published by Search vice-president Liz Reid on 6 August 2025, is that total organic click volume has been relatively stable year over year, that average click quality has risen, and that third-party reports of dramatic declines rest on flawed methods. Google defines a quality click as one where the user does not quickly click back.

Both statements can hold at once. Pew measured yield per search. Google measured aggregate volume across the whole web. If searches grow fast enough, per-search yield can halve while total clicks hold flat. Alphabet’s filings support the growth half. Alphabet’s Q2 2026 earnings release lodged with the Securities and Exchange Commission reports Google Search and other revenues of $63,271 million against $54,190 million a year earlier, a rise of 16.8 per cent. Meanwhile the Reuters Institute for the Study of Journalism reported in June 2026 that Google organic traffic to more than 2,500 sites was down by a third globally between November 2024 and November 2025.

An SME spends clicks, and clicks are what moved.

Two-thirds of last year’s earning pages went dark

The stage model’s other load-bearing assumption is compounding. Publish consistently, the advice runs, and the archive keeps paying. That claim is testable: cut the record into two equal windows and follow the same URLs across the join.

On the same property, 344 URLs drew impressions between 1 June and 31 December 2025, and 230 of them earned at least one click. In the following seven months, 1 January to 31 July 2026, only 85 of those 230 earned a click again. Sixty-three per cent stopped earning inside seven months, taking 191 clicks to zero.

Now the evidence that cuts against the thesis. The 85 survivors did compound. They earned 689 clicks in the first window and 1,105 in the second, a rise of 60 per cent with no new publishing behind it. Compounding runs through 37 per cent of the pages that ever earned a click here, and the remaining 63 per cent turned into maintenance that a content calendar kept adding to.

The same shape appears on the two other properties, one of which is the site carrying this article. On that one, 414 URLs drew impressions across twelve months and 386, or 93 per cent, never converted a single impression into a click, while absorbing 37 per cent of the impressions its pages drew. On the third, 252 of 301 URLs did the same. Between them, 638 URLs drew Google impressions for a year and returned nothing for the cost of hosting them.

AI assistants sent 80 sessions, 2.4 per cent of organic search

If discovery is migrating to answer engines, it should show in the analytics. On the largest of the three properties, Google Analytics 4 recorded 3,368 organic search sessions over the twelve months to 31 July 2026. Referrals from the four largest AI assistant domains totalled 80 sessions, of which 73 came from a single engine. That is 2.4 per cent of organic search sessions, and it is the entire channel.

Restructuring an SME’s publishing around answer engines today spends real money against 80 sessions. And the quality question stays open in both directions. Google claims clicks from result pages carrying AI Overviews are higher quality, with users staying longer, while the Reuters Institute found that 42 per cent of chatbot news users say they always or often click through to the original source. Both numbers are too small to justify a rebuild and big enough to justify a monthly line.

Switzerland runs ten monthly searches for organic growth strategy

Google Ads Keyword Planner returns an average of ten monthly searches in Switzerland for the phrase “organic growth strategy” over the twelve months to July 2026, and 90 across Switzerland, France, the United Kingdom and the United States combined. The phrase “organic growth engine”, the shape of a thousand blog posts, returns no measurable volume. Keyword Planner rounds its volumes and groups close variants, so treat those as orders of magnitude rather than counts.

An SME publishing weekly into a market that generates ten searches a month for its core topic is running a factory for a demand curve of ten searches a month. Kainjoo argued the budget half of this before, in the case for buying less demand and compounding more of it. The measurement half is what Search Console adds.

Google’s guidance points the same way. Google’s generative-AI optimisation guide, updated in July 2026, says creating content people find unique, compelling and useful will likely influence a site’s presence in generative AI search more than anything else in the document, and defines commodity content as common knowledge that could have come from anyone. Google’s helpful-content guidance asks whether a page carries original information, reporting, research or analysis. Publishing cadence appears nowhere in either.

Run these four checks before publishing anything else

Organic still pays. Audit the engine before feeding it again. Four checks, each an afternoon with one export.

Split branded vs non-branded search and compute the two CTRs separately. Export twelve months of query data, tag every query containing the company or product name, and total each group. Tag the misspellings too, because two carried 558 impressions and 18 clicks here. If the branded group carries most of the clicks, the brand is doing the acquiring and the content is doing something else, whatever the traffic chart implies. A high branded share is neither good nor bad on its own. It is good if the brand is growing and bad if it is the only thing working, and the non-brand CTR tells you which.

Find the URLs that carry 80 per cent of the clicks, and count them. On the property examined here, the top five URLs took 53 per cent of a seven-month click total and the top twenty took 70 per cent. Everything outside that set costs review time and internal-link equity, returning close to nothing.

Run the two-window survivor test. Cut the last fourteen months in half, list the URLs that earned a click in the first window, and check which earned again in the second. Refresh the survivors, the pages that actually compound. Consolidate or retire the rest instead of adding siblings alongside them.

Count the impressions your zero-click URLs threw away. A third of the impressions on the second property landed on pages that never converted one. That is reach already won and wasted at the title or the intent match, cheaper to fix than to earn again. Treat it as a routing problem, the same discipline Kainjoo applies when leads die in the handoff, because a click lost at the snippet and a lead lost at the handoff are one leak at two ends.

Those four checks make the stage model run in the order the evidence supports. Earn the name first, because branded queries are the only ones converting at double digits. Then protect the small survivor set, which is the part that compounds, and publish against a topic only where measured demand justifies the mandate. Pupsic runs that audit as the opening move of a RevOps mandate, and it usually ends with fewer articles and better margin.

References

1. Pew Research Center. Google users are less likely to click on links when an AI summary appears in the results. 22 July 2025. https://www.pewresearch.org/short-reads/2025/07/22/google-users-are-less-likely-to-click-on-links-when-an-ai-summary-appears-in-the-results/
2. Reid, L. Google. AI in Search is driving more queries and higher quality clicks. 6 August 2025. https://blog.google/products-and-platforms/products/search/ai-search-driving-more-queries-higher-quality-clicks/
3. Alphabet Inc. Alphabet Announces Second Quarter 2026 Results, Exhibit 99.1 filed with the United States Securities and Exchange Commission, 22 July 2026. https://www.sec.gov/Archives/edgar/data/1652044/000165204426000066/googexhibit991q22026.htm
4. Reuters Institute for the Study of Journalism, University of Oxford. Digital News Report 2026, overview and key findings. 16 June 2026. https://reutersinstitute.politics.ox.ac.uk/digital-news-report/2026/dnr-executive-summary
5. Court, D., Elzinga, D., Mulder, S. and Vetvik, O. J. The consumer decision journey. McKinsey Quarterly, 1 June 2009. https://www.mckinsey.com/capabilities/growth-marketing-and-sales/our-insights/the-consumer-decision-journey
6. Google Search Central. AI Features and Your Website. Last updated 10 December 2025. https://developers.google.com/search/docs/appearance/ai-features
7. Google Search Console Help. Performance report (Search), how clicks and impressions are counted by search type. https://support.google.com/webmasters/answer/7042828
8. Google Search Console Help. Performance reports, anonymized queries and data aggregation. https://support.google.com/webmasters/answer/7576553
9. Google Search Central. Optimizing your website for generative AI features on Google Search. Last updated 10 July 2026. https://developers.google.com/search/docs/fundamentals/ai-optimization-guide
10. Google Search Central. Creating helpful, reliable, people-first content. Last updated 10 December 2025. https://developers.google.com/search/docs/fundamentals/creating-helpful-content
11. Google Search Console API. Search Analytics: query, request limits and dimensions. Last updated 11 August 2026. https://developers.google.com/webmaster-tools/v1/searchanalytics/query
12. Google Ads Help. About Keyword Planner forecasts and average monthly searches. https://support.google.com/google-ads/answer/3022575
13. Fust, A. and Graf, A. Les 99% de l’economie suisse. La Vie economique, 17 April 2023, citing the Federal Statistical Office structural business statistics (STATENT). https://dievolkswirtschaft.ch/fr/2023/04/les-99-de-leconomie-suisse/
14. State Secretariat for Economic Affairs (SECO). Chiffres sur les PME, entreprises et emplois, STATENT 2023. https://www.kmu.admin.ch/fr/chiffres-sur-les-pme-entreprises-et-emplois
15. Pupsic. Demand generation for Swiss SMEs, buy less, compound more. https://pupsic.ch/demand-generation-for-swiss-smes-buy-less-compound-more/
16. Pupsic. Swiss SMEs Lose Most Leads in the Handoff, Not the Pitch. https://pupsic.ch/revops-switzerland-lead-routing-speed-to-lead/
17. Pupsic. RevOps services. https://pupsic.ch/services/
18. Kainjoo first-party data. Google Search Console Search Analytics API and Google Analytics 4 exports for three Swiss B2B properties under Kainjoo management, including Kainjoo’s own group domain, which is the largest of the three, pupsic.ch, which carries this article, and treatmybrand.com; 1 August 2025 to 31 July 2026. Google Ads Keyword Planner historical metrics, same period. Cantons and identifying queries withheld.