Pogosticking: The Click Pattern That Quietly Decides Who Ranks

You know the page. You searched a question, clicked a promising headline, and landed on a wall. A hero image the size of a billboard, no actual words in sight. You scroll. A newsletter popup. You hit the X. You scroll again. A second popup, this one wants you to register for "three free articles." You hit the X again. You scroll past a wall of social buttons and finally reach paragraph one, which turns out to be three sentences before it dissolves into a "subscribe to keep reading" gradient. So you do what most of us do. You hit the back button and you click the next result.

That move you just made has a name. It is called pogosticking, and it is one of the most quietly decisive things that happens in search. This is the story of what it is, why it matters far more than Google admitted for years, and why I watched a whole industry walk straight into the ocean because nobody upstairs wanted to hear it.

What pogosticking actually is

Pogosticking is the back-and-forth bounce between the search results and the pages they point to. A user clicks result one, comes straight back, clicks result two, comes back again, clicks result three, and this time stays. To Google, that final stop is loud. It says: results one and two did not answer the question. Result three did.

People confuse this with bounce rate, and that confusion lets a lot of bad SEO advice survive. Bounce rate is a Google Analytics number about whether someone triggered a second event on your site. Pogosticking is something else entirely. It is Google watching its own search results, from its own click stream, in Search and in Chrome, and noticing that you were a pit stop on the way to a better answer. You do not control the tool measuring it, you cannot game it with a fake scroll script, and your analytics dashboard cannot even see it.

For years they said clicks did not count. Then they testified

Here is the part that still pisses me off a little. For the better part of a decade, Google representatives waved this away. Clicks are too noisy, they said. We do not use engagement as a ranking signal, they said. A generation of SEOs got called tinfoil-hat conspiracy theorists for insisting that user behavior obviously mattered.

Then two things happened. In the 2023 and 2024 U.S. Department of Justice antitrust trial, Google's own VP of Search, Pandu Nayak, described a system called NavBoost under oath, and called it one of Google's most important signals. Then in May 2024, an enormous leak of Google's internal Search documentation spilled the actual field names. And there they were, in black and white: goodClicks, badClicks, and lastLongestClicks.

Read those three names again, because they describe pogosticking precisely. A good click is one where the user stays. A bad click is one where they bounce back to the results, which is to say, a pogostick. And the last longest click is the kill shot: in a search session where someone tried several results, the one they stayed on longest before they stopped searching gets a strong positive signal. NavBoost re-ranks results based on thirteen months of this aggregated behavior. It does not produce the first ranking. It watches what humans actually do with that ranking, and then it reshuffles. The competitor who held your visitor just took your spot, and you handed it to them.

So when someone tells you "Google says bounce rate is not a ranking factor," they are technically correct and completely missing the point. Google does not need your Analytics bounce rate. It has something far better: the searcher coming back to Google and being happier with the next guy.

The newsroom that would not listen

I spent years doing technical SEO inside a large American newspaper group, and I will tell you the single hill I died on more times than any other. User experience is number one on Google's list. Not because Google is a benevolent monk who wants the web to be lovely. Because a satisfied searcher comes back to Google tomorrow and clicks another ad. Google's entire business depends on you finding what you needed, fast, so you trust the box and return to it. Every UX rule they publish is downstream of that one commercial fact. Make the human happy, and Google makes money. That is the deal underneath all of it.

And no, before anyone calls me a Google fanboy: I am not. Not even close. I have degoogled my own life about as far as a person reasonably can, a GrapheneOS phone, iOS where it earns its place, even the Android TV in my living room is rooted and stripped down to the studs. I do not trust their practices and I will not pretend to. Which is precisely why you should believe me when I say UX is their self-interest and not their virtue. This is not Google PR talking, it is a critic explaining how the machine actually pays itself.

You want the tell? In 2018, Google quietly lifted "don't be evil" out of the opening of its Code of Conduct, where it had sat for almost two decades, and tucked it into the very last line. Half the internet screamed that Google had deleted it, and that is the part worth getting right, because they had not. It was demoted, not removed, shuffled from the mission statement to a footnote you only reach if you read all the way to the bottom. The timeline is well documented, and you can still find the line at the foot of the current Code of Conduct if you scroll far enough. Call that whatever you want. My point is narrower, and it is the one that bugs me when I see the lazy version online: do not be ignorant about it. The line is not gone, it is just no longer the first thing they want you to read.

I would sit in those meetings and explain it plainly. This article cannot even be seen by Google, because the top of the page is one enormous image with no real text, and then the words that do exist are locked behind a registration wall before a crawler or a reader can reach them. The pogosticking on these templates is alarming. People are bouncing in two seconds. We are training Google to think we never answer anything.

And the answer, more often than not, was a kind of polite institutional arrogance. Not from the journalists, to be clear. Most of the writers just wanted to be read, and a few of them quietly agreed with me in the hallway. It came from higher up, from the belief that everything the outlet published was a Pulitzer contender and therefore belonged behind a golden wall, untouchable, un-scrollable, un-skimmable. The idea that a page had to actually serve the reader in the first three seconds felt almost insulting to that worldview. Putain, the conversations were exhausting. And this is the quiet absurdity of doing SEO inside a publisher: you fight two wars at once, one against every competitor in the results, and a stranger one against your own organization, which keeps doing the exact things that guarantee you lose the first war and then asks you to rank anyway. I understood the business constraint. Newsrooms need revenue, real revenue, and the ad market had been gutted. I never pretended otherwise. But there is a humane way to ask someone to pay, and there is the hostile way, and they kept choosing hostile and then blaming the algorithm.

Then the reckoning came, and it had a name too

I left that industry. And I do not have to imagine what happened next, because we can all see it. AI Overviews, ChatGPT, Perplexity, the answer engines. The model reads the open web, synthesizes the answer, and hands it to the user who never clicks through at all. The publishers who spent fifteen years teaching readers that clicking their link meant a fight with three popups and a paywall had already built the muscle memory. Now a machine offers the same answer with zero friction, and the back button does not even get involved.

Here is the quiet symmetry. The pogostick was the market telling those publishers the truth in advance. A lot of those two-second bounces were a reader saying "this is not worth the hassle." They had years of that signal screaming at them, and they read it as an algorithm problem instead of a respect problem. The sites that survived the AI gut-punch are, overwhelmingly, the ones that earned the long click. The ones that let you read, that answered fast, that treated the reader like a guest instead of a mark.

Google had to make a rule, and they still did not get the message

Sit with the absurdity for a second. Intrusive interstitials, the popups that bury the content the instant you arrive, are such a reliably hostile pattern that Google built an actual penalty for them back in January 2017. A search engine, whose entire job is to point people at useful pages, looked at how publishers were treating arriving visitors and decided it had to step in and dock them for it. That is not a subtle hint. That is the referee stopping the game to tell you to quit kicking the other guy.

And they are still doing it. Still serving the full-screen register-wall, still convinced that this is the year readers finally roll over and accept it. Oh merde, the optimism. Meanwhile the stakeholders who sign off on these templates sit in a room being told the subscriber funnel is converting nicely, soothed by their own lullaby, comfortably disconnected from the actual experience of being a human who just wanted to read one article without a fight.

The reader revolt already has a URL

You want proof the market rejected this? People built an entire counter-infrastructure to route around it. Archive.today, also reachable as archive.ph and archive.is, exists in large part because readers got tired of the wall and just wanted the words. Smry.ai and a parade of others do the same job. 12ft.io did it too, until the News Media Alliance got it taken down in July 2025, and here is the tell: killing the tool changed nothing, because the demand never moved. A replacement was up almost immediately. You cannot DMCA your way out of being annoying.

Then look at the etiquette that grew up around it. In huge corners of Reddit it is considered flat-out rude to post a hard-paywalled link without dropping a free mirror in the comments. Read that again, because the inversion is the entire point: the community treats serving the reader a paywall as the impolite act. The wall is the faux pas. How fucking dare you make me fight to read this, is the prevailing mood, and the person who posts the clean version is the hero of the thread. That is your audience telling you, in public, in their own house, exactly how they feel about the door you built.

And, in my opinion, stop deleting your own archive

Here is one that genuinely hurts to watch, and I will flag it up front as my own bias, not gospel. Publishers delete old pages. Not archive them, delete them, and let the URLs rot into 404s. I get the instinct. When you publish somewhere between five and a thousand articles a day, you obviously cannot keep them all one click from the homepage, and nobody is asking you to. Subcategories get crowded, campaigns end, sections get retired. Fine. But there is a canyon between gracefully demoting an old article and erasing it.

Let me own the bias completely: I am a huge fan of the New York Times archive. It is, frankly, gorgeous. Clean, restrained, beautifully designed, and almost every article is still reachable and still linked. They treat their back catalog like the asset it is, decades of work that still earns clicks, still earns citations, still earns trust. That is what a real archive looks like, and it is not an accident. It is a decision to respect the reader who shows up three years late.

Now picture the opposite, which you have all lived. You search something, you find the perfect headline, you click, and you get the site banner, the nav band, and a 404 staring back. In my book that is an automatic negative point. You came to read, the door was bricked over, and you are gone in under a second, which, if you have been paying attention, is the textbook pogostick. Every deleted page that still had inbound links or search demand is a page you turned into a tiny bounce machine on purpose. You cannot ask a search engine to value your domain while you are busy demolishing the parts of it people still want. This is the same self-inflicted wound as broken internal links, just with the dynamite set off by your own hand.

And again, this is my read, so take it as that. Demote, consolidate, merge thin pieces into stronger ones, redirect the genuinely dead to the nearest living relative. All good, all healthy. Just stop nuking content into hard 404s and then wondering why the archive that used to bring traffic went quiet.

How not to be a pogostick machine

None of this is mystical. If you want the long click instead of the bounce, you make the page worth staying on, and you make that obvious immediately. Concretely:

Answer the query above the fold. The thing they searched for should be visibly addressed in the first screen, in real text, before any image, popup, or pitch. Match the intent of the query, do not bait it.

Put real, crawlable text near the top. A giant hero image with the actual content baked into it, or hidden behind a script, is invisible to Google and to AI crawlers alike. If a machine cannot read your answer, it cannot rank it or cite it.

Do not assault people on arrival. Intrusive interstitials that cover the content the moment someone lands are the single fastest way to manufacture a bounce. Delay them, shrink them, or kill them. Monetize after you have delivered value, not before.

Be fast, especially on the phone. A slow page is a bounce in waiting. The user is gone before your hero even paints. Speed and interaction responsiveness are pogosticking insurance.

Match the promise of your title. Clickbait wins the click and loses the session. The gap between what your snippet promised and what the page delivers is measured in exactly the bounce NavBoost is built to catch. Write the title the page can actually keep.

You will notice none of this is a trick. You cannot fake the long click, because the thing being measured is whether a real human got what they came for. That is the entire point, and it is also the good news: the work that satisfies a reader is the same work that satisfies the algorithm, because the algorithm is just a proxy for the reader.

The part nobody upstairs wanted to hear

User experience was never the soft, optional, nice-to-have line item at the bottom of the SEO deck. It is far closer to the heart of the work than most decks admit. Google rewards it because Google's money depends on it. The AI engines reward it because their usefulness depends on it. And readers reward it because, shockingly, people remember who respected their time.

The pogostick was the warning, sent for free, for years. The shame is not that the publishers got disrupted. The shame is that the signal was sitting right there in the bounce, and the arrogance was too loud to hear it. Do not make the same mistake. Earn the long click, and you are building the one thing both Google and the machines actually reward: a page a human is glad they stayed on.

So here is the part that survives every algorithm update, every leak, and every shiny new model. You can stack every rich schema markup ever written, sprinkle in flawless semantic keywords, architect a perfect internal linking structure, ship next-gen images and a board of green Core Web Vitals, and it buys you nothing if you screw the visitor at the door. None of it outranks a human deciding you were not worth the hassle. You are not unique. Whatever you are serving, someone else serves it without the ambush, and they sit one back-button away. So burn this into the wall above your desk: you are not unique, and a click is not owed to you. A click is a privilege, handed over by a person who can revoke it in two seconds, and the whole game, the only game, is being worth the one they gave you.