---
title: "The Click Memory — how a click becomes the memory that ranks the next person"
url: https://martech.llc/research/the-click-memory
publishedAt: 2026-06-27
updatedAt: 2026-06-27
author: sundar
category: research-note
summary: "For years Google denied clicks ranked pages. A US antitrust trial, a granted-patent family, and a leaked internal API ended the debate. This maps the Click Memory — the seven layers that turn user behaviour into rank, and how AI answers inherit that memory even as they starve it."
soWhat: "Clicks are an adjudicated ranking signal held in a rolling ~13-month memory — so win the satisfied long click, because AI answers inherit that memory even as they thin it out."
tags: ["search-engine-optimization","generative-engine-optimization","ai-search","behavioral-ranking","navboost"]
keywords: ["does google use clicks for ranking","navboost ranking signal","good clicks bad clicks","click signals seo","do clicks affect ai overviews","google click data ranking","dwell time ranking factor","behavioral signals generative engine optimization"]
claims: [{"id":"claim-1","text":"An internal Google training deck admitted at the US v. Google trial names 'The 3 Pillars of Ranking' as body, anchors, and user-interactions, stating that user-interactions include clicks, attention on a result, swipes on carousels and entering a new query.","source":"https://www.justice.gov/d9/2023-11/417508.pdf","sourceTitle":"Trial Exhibit UPX0004 — 'Life of a Click (user-interaction)', US v. Google","sourceDate":"2023-11-01"},{"id":"claim-2","text":"Under oath, Google's VP of Search confirmed that the Navboost system memorizes user clicks for every query received in the prior 13 months.","source":"https://thecapitolforum.com/wp-content/uploads/2023/10/101823-USA-v-Google-PM.pdf","sourceTitle":"Cross-examination of Pandu Nayak, US v. Google bench-trial transcript (Day 24 PM)","sourceDate":"2023-10-18"},{"id":"claim-3","text":"The US v. Google liability opinion found that learning from user click feedback has been perhaps the central way that web ranking has improved for 15 years.","source":"https://caselaw.findlaw.com/court/us-dis-crt-dis-col/116454429.html","sourceTitle":"United States v. Google LLC — Memorandum Opinion (liability), Judge Amit P. Mehta","sourceDate":"2024-08-05"},{"id":"claim-4","text":"The 2024 leaked Google Content Warehouse API documentation contains explicit Navboost click fields — bad clicks, good clicks, last longest clicks, unsquashed clicks and unsquashed last longest clicks.","source":"https://ipullrank.com/google-algo-leak","sourceTitle":"Secrets from the Algorithm: Google Search's Internal Engineering Documentation Has Leaked — iPullRank (Mike King)","sourceDate":"2024-05-27"},{"id":"claim-5","text":"The first-hand analysis of the leak cautioned that the documentation does not show the weight of particular elements in the search ranking algorithm, nor prove which elements are used in the ranking systems.","source":"https://sparktoro.com/blog/an-anonymous-source-shared-thousands-of-leaked-google-search-api-documents-with-me-everyone-in-seo-should-see-them/","sourceTitle":"An Anonymous Source Shared Thousands of Leaked Google Search API Documents with Me — SparkToro (Rand Fishkin)","sourceDate":"2024-05-27"},{"id":"claim-6","text":"A granted Google patent describes tracking user click data and transforming it into a 'click fraction' used to re-rank future results, distinguishing long clicks from short clicks as a proxy for satisfaction.","source":"https://patents.google.com/patent/US8661029B1/en","sourceTitle":"US8,661,029 B1 — Modifying search result ranking based on implicit user feedback (Google)","sourceDate":"2014-02-25"},{"id":"claim-7","text":"A separate granted Google patent describes adjusting a document's relevance based on a 'temporal element of the user feedback', down-weighting aged selections — the mechanism behind decay of click signals.","source":"https://patents.google.com/patent/US9092510B1/en","sourceTitle":"US9,092,510 B1 — Modifying search result ranking based on a temporal element of user feedback (Google)","sourceDate":"2015-07-28"},{"id":"claim-8","text":"The cascade model — where users view results from top to bottom and leave as soon as they see a worthwhile document — is the best explanation for position bias in early ranks.","source":"https://www.microsoft.com/en-us/research/publication/an-experimental-comparison-of-click-position-bias-models/","sourceTitle":"An Experimental Comparison of Click Position-Bias Models — Microsoft Research (WSDM 2008)","sourceDate":"2008-02-01"},{"id":"claim-9","text":"The Dynamic Bayesian Network click model separates a result's attractiveness, which drives the click, from its actual relevance, which governs whether the user stays satisfied after clicking.","source":"https://dl.acm.org/doi/10.1145/1526709.1526711","sourceTitle":"A Dynamic Bayesian Network Click Model for Web Search Ranking — Chapelle & Zhang (WWW 2009)","sourceDate":"2009-04-20"},{"id":"claim-10","text":"Microsoft openly documents that Bing uses engagement signals — including whether users spent time on a clicked result or quickly returned to Bing — as inputs to ranking.","source":"https://support.microsoft.com/en-us/bing/how-bing-delivers-search-results","sourceTitle":"How Bing delivers search results — Microsoft","sourceDate":"2025-03-01"},{"id":"claim-11","text":"Google states that its generative AI features on Search are rooted in its core Search ranking and quality systems, so the best practices for SEO continue to be relevant.","source":"https://developers.google.com/search/docs/fundamentals/ai-optimization-guide","sourceTitle":"Top ways to ensure your content performs well in Google's AI experiences — Google Search Central","sourceDate":"2026-06-15"},{"id":"claim-12","text":"On a behaviourally-tracked panel, the presence of an AI Overview cut the rate of clicking a traditional organic result from 15% to 8%, while only 1% of users clicked a link inside the summary.","source":"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/","sourceTitle":"Google users are less likely to click on links when an AI summary appears — Pew Research Center","sourceDate":"2025-07-22"},{"id":"claim-13","text":"On a tens-of-millions desktop clickstream panel, fewer than half of US Google searches in March 2025 ended in an organic click — 40.3% organic click versus 27.2% ending with no click at all.","source":"https://23904045.fs1.hubspotusercontent-na1.net/hubfs/23904045/EN%20Datos%20State%20of%20Search%20report%20Q1%202025.pdf","sourceTitle":"State of Search Q1 2025 — Datos (a Semrush company) × SparkToro","sourceDate":"2025-04-01"},{"id":"claim-14","text":"A granted Google patent computes a site quality score from user-demand signals — a ratio built from the count of reference (brand-seeking) queries to the count of queries associated with the site through user selection of its results.","source":"https://patents.google.com/patent/US9031929B1/en","sourceTitle":"US9,031,929 B1 — Site quality score (Google; co-invented by Navneet Panda)","sourceDate":"2015-05-12"},{"id":"claim-15","text":"Google's public roster of named ranking systems lists systems such as BERT, RankBrain, and PageRank but names no click-, engagement-, or behavioural-signal system at all.","source":"https://developers.google.com/search/docs/appearance/ranking-systems-guide","sourceTitle":"A Guide to Google Search Ranking Systems — Google Search Central","sourceDate":"2025-12-10"}]
---
# The Click Memory — how a click becomes the memory that ranks the next person
Google's ranking does not just read your page. It remembers what real users did the last time it showed them pages like yours. A US antitrust trial, a granted-patent family dating to 2006, and a leaked internal ranking API now converge on the same mechanism: a behavioural memory that turns clicks into rank. This is the Click Memory — and AI answers inherit it.
For most of the last decade, "do clicks affect rankings?" was the question Google answered most evasively. Representatives downplayed it; the SEO industry argued about it; nobody outside Mountain View could see the machinery. Then the machinery was entered into evidence. This piece does something the click debate never could before 2023: it builds the model from primary sources you can open yourself — court exhibits, sworn testimony, a federal opinion, granted patents, peer-reviewed click models, and the platform documentation — and it closes by drawing the honest line none of that evidence crosses.
## Did the click debate actually end?
It did, and the place it ended was a courtroom. In *United States v. Google*, the government entered an internal Google training deck titled "Life of a Click." That exhibit lays out what Google internally calls [the three pillars of ranking](https://www.justice.gov/d9/2023-11/417508.pdf) — body (what the document says about itself), anchors (what the web says about it), and user-interactions (what users say about it) — and notes that user-interactions include "clicks, attention on a result, swipes on carousels and entering a new query." One of the three pillars is, in Google's own words, behaviour.
Then it was confirmed under oath. Google's VP of Search testified that [the Navboost system memorizes past clicks for past queries](https://thecapitolforum.com/wp-content/uploads/2023/10/101823-USA-v-Google-PM.pdf), is trained on user data, and retains that click information for every query received in the prior 13 months. And the court itself made it a finding of fact. Judge Amit Mehta's [liability opinion](https://caselaw.findlaw.com/court/us-dis-crt-dis-col/116454429.html) concluded that ranking relies heavily on user click-and-query data and that "learning from this user feedback is perhaps the central way that web ranking has improved for 15 years."
> Learning from this user feedback is perhaps the central way that web ranking has improved for 15 years.
A 17-year Google search-quality engineer put the years of public hedging more bluntly in his own testimony, [as reported by Search Engine Land](https://searchengineland.com/former-googler-google-using-clicks-in-rankings-432401): the debate was never whether clicks were used, but why anyone pretended otherwise.
> Pretty much everyone knows we're using clicks in rankings.
This matters because everything downstream — the patents, the leak, the AI-answer question — is no longer a theory in search of confirmation. The confirmation came first, from the witness stand. The rest of this piece is the mechanism behind it.
## What is the Click Memory?
The figure above is the whole argument in one diagram, so read it as a pipeline rather than a list. A search enters at the top as an *examination* — whether a result is even seen — and exits at the bottom as an entry in a memory that will shape the next person's results. Each layer is independently sourced, and each exposes exactly one lever an operator can actually pull. The crucial reframing: the engine is not scoring your page in isolation. It is replaying aggregate behaviour and promoting whatever satisfied people last time.
The reason this is worth a framework rather than a tip is that the layers interact. Winning a click you can't satisfy (L1 without L2) actively teaches the memory to prefer a rival. Earning satisfied clicks on desktop while neglecting mobile (L2 without L3 consistency) banks signal in a partition that may not be the one a given searcher draws from. The layers are a system, and the system has a direction of flow.
## How does the system tell a good click from a bad one?
This is the layer everything else rests on, because not all clicks count the same. The distinction is between a click that is *kept* and a click that is *regretted*. A granted Google patent describes tracking click data and transforming it into a ["click fraction"](https://patents.google.com/patent/US8661029B1/en) used to re-rank future results, explicitly separating long clicks — longer views that imply satisfaction — from short ones. The 2024 leak gave that distinction its production vocabulary. The leaked [Content Warehouse API documentation](https://ipullrank.com/google-algo-leak) lists explicit click fields — "bad clicks, good clicks, last longest clicks, unsquashed clicks, and unsquashed last longest clicks."
> bad clicks, good clicks, last longest clicks, unsquashed clicks, and unsquashed last longest clicks are all considered
Notice the phrase *last longest click*. The signal isn't merely that you were clicked; it's that yours was the click the user stayed on longest before the session ended — the result that resolved the query. This is not unique to Google. Microsoft openly documents that Bing uses engagement signals, asking directly whether ["users spend time on these search results they clicked through or quickly return to Bing."](https://support.microsoft.com/en-us/bing/how-bing-delivers-search-results)
The operator consequence is uncomfortable for anyone optimising for click-through rate alone: ranking for a query you don't satisfy is worse than not ranking for it. Every pogo-stick back to the results page is a vote for the next result down. The behavioural layer rewards the answer, not the headline.
## Why does your snippet have to beat your rank?
Before any of that, a click has to be *possible*, and attention is not distributed evenly down the page. The foundational study here established that a ["cascade" model](https://www.microsoft.com/en-us/research/publication/an-experimental-comparison-of-click-position-bias-models/) — where users read results top to bottom and stop as soon as they find a worthwhile one — best explains why click probability falls with rank. Lower positions are examined less, so they are clicked less, regardless of how good they are.
Raw click-through, then, is contaminated by position. A serious ranking system has to correct for it before trusting clicks, and the literature is explicit about how. The [Dynamic Bayesian Network click model](https://dl.acm.org/doi/10.1145/1526709.1526711) separates a result's *attractiveness* — which drives the click — from its *actual relevance*, which governs whether the user stays satisfied afterward. Google's own patents describe modelling presentation bias to recover an unbiased relevance estimate from clicks, and its researchers have published [production-scale methods for estimating that bias from ordinary click logs](https://marc.najork.org/papers/wsdm2018.pdf) without degrading the user experience.
For the operator, debiasing cuts both ways. You are corrected *against* at every rank below the top — the system knows you got fewer clicks partly because you sat lower, and it won't fully credit you for that. So your title and snippet have to over-perform your position: earn more examination and more satisfied clicks than your rank would predict, and you give the debiasing math a reason to lift you.
## The memory has a window — and that is your opening
A memory that never forgot would ossify; the incumbent that won in 2015 would win forever. It doesn't, because the table is time-bounded. The same testimony that confirmed Navboost confirmed its horizon: clicks are retained on a rolling window of about 13 months. And a granted Google patent describes [adjusting relevance by a "temporal element" of user feedback](https://patents.google.com/patent/US9092510B1/en), down-weighting aged selections so that recent behaviour counts for more.A click is a deposit, and deposits decay. The page that stops earning satisfied clicks doesn't hold its rank — it slides out of the favourable memory while a fresher answer slides in.
This is the most actionable property in the whole model. Decay means the leaderboard is always partly open. A challenger who starts earning the long click *now* is writing into a window that will be the active one in a few months; an incumbent coasting on a page that no longer satisfies is watching its deposit age out. Refreshing a page is not housekeeping — it is re-depositing into the memory that ranks.
## Does brand demand really move sitewide rank?
The layers so far operate per page. The last behavioural layer operates per *site*, and it runs on a kind of click most teams ignore: the navigational query, where someone searches for you by name. A granted Google patent computes a [site quality score](https://patents.google.com/patent/US9031929B1/en) from user-demand signals — a ratio built from the count of reference, brand-seeking queries against the count of queries the site is merely associated with through result selection. Co-invented by the engineer the "Panda" quality system is named after, it encodes a simple intuition: a site people seek out by name is, on the evidence of their behaviour, a site that satisfies.
The lever is the hardest and the most durable: manufacture genuine brand demand. Every person who searches your name instead of your category is casting the highest-value behavioural vote there is — and it rolls up into a sitewide modifier that lifts every page you publish, not just the one they were looking for.
## What does this mean for AI answers?
Here the evidence gets more interesting and more honest at once. The behavioural memory was built for the ten blue links. Does it reach the AI answer? Google's documentation says the plumbing is shared. Google states that its [generative AI features are "rooted in" its core Search ranking and quality systems](https://developers.google.com/search/docs/fundamentals/ai-optimization-guide), and that grounding retrieves pages through those same core systems. If core ranking decides what gets retrieved, and behaviour shapes core ranking, then the Click Memory governs whether you are even eligible to be cited.
But the AI answer also *starves* the memory it inherits, and the data on that is stark. A [Pew Research Center clickstream panel](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/) found that when an AI summary appeared, the rate of clicking a traditional result fell from 15% to 8%, and just 1% of users clicked a link inside the summary itself. The longer trend is the same direction. On a tens-of-millions desktop panel, [fewer than half of US Google searches in March 2025 ended in an organic click](https://23904045.fs1.hubspotusercontent-na1.net/hubfs/23904045/EN%20Datos%20State%20of%20Search%20report%20Q1%202025.pdf) — 40.3% organic click against 27.2% ending with no click at all.
That is the paradox an operator has to hold: AI answers read from the click memory at retrieval but contribute almost nothing back to it. Academic work has started naming the disconnect — [generative AI search collects only coarse feedback on the final answer](https://arxiv.org/abs/2505.14680), not the fine-grained per-document clicks that taught web ranking for two decades. The behavioural signal that historically self-corrected ranking is thinning exactly where the answers are moving. The window to be remembered as the satisfying source is closing, not widening.
## What the evidence does NOT prove
A model this clean invites overreach, so here is the discipline. None of these sources publishes the *weights*. The first-hand analysis of the leak was emphatic that the [documentation "doesn't show things like the weight of particular elements in the search ranking algorithm, nor does it prove which elements are used in the ranking systems."](https://sparktoro.com/blog/an-anonymous-source-shared-thousands-of-leaked-google-search-api-documents-with-me-everyone-in-seo-should-see-them/) The court redacted the formulas. The patents are *methods*, not proof of the live spec. And Google's public posture still omits the mechanism entirely: its [official roster of named ranking systems](https://developers.google.com/search/docs/appearance/ranking-systems-guide) lists BERT, RankBrain, and PageRank — and names no click or engagement system at all.
Two honest caveats sit on top. First, the trial adjudicated *core web ranking*; the claim that the AI synthesis layer re-ranks on clicks is inferred from Google's "rooted in core ranking" documentation, not stated in the opinion. Second, the witness who confirmed Navboost also cautioned it is "a factor … by no means the only factor," and that many documents have no clicks at all. The Click Memory is a defensible model of *what is measured*. It is not a claim about *how much each signal is worth* — and any vendor who tells you they know the weights is selling you the part nobody outside Google can see.
## Run one query through the Click Memory
Models are easiest to trust when you watch them run on something concrete. Take a single commercial query, a single page sitting at rank four, and walk it through all the layers — what each one tests, and the one lever that wins it.
The pattern the walkthrough exposes is the whole strategy in miniature. The first three layers are a single session you can win this week: a sharper snippet, an honest promise, an answer above the fold. The last layers are the moat — a *pattern* of satisfied clicks the system memorizes, keeps only while you keep earning it, and increasingly hands to the AI answer at retrieval. Behavioural rank compounds for whoever satisfies people consistently, and decays for whoever coasts. That is the entire game, and it has been the game, on the court's own finding, for fifteen years.