Pogo-Sticking Explained: The SEO Signal Google Uses to Decide Which Results Are Actually Useful

Pogo-sticking, in the strict SEO sense, is a user behavior pattern with three parts that must all occur in sequence: (1) a user clicks a result on the Google SERP, (2) spends a relatively short interval on the destination page, and (3) returns to the same SERP and clicks a different result for the same query. All three parts are required. A user who clicks a result, stays for ten minutes, then closes the tab hasn't pogo-sticked. A user who opens five results in background tabs has not pogo-sticked. A user who returns to Google with a new query has not pogo-sticked — that's query refinement, a different signal entirely.

The "short interval" threshold is not a fixed number. Internal Google documents and patent language suggest a dynamic threshold tuned per query class — informational queries tolerate longer dwells before reading as pogo-stick, transactional queries expect faster satisfaction. The working number most SEO teams use is under 25 seconds for informational and under 15 seconds for transactional, but these are approximations. What matters is the relative comparison: a user who spent shorter time on result A than the average session on the same URL across the same query is pogo-sticking.

The term is sometimes used loosely to mean "user returned to the SERP quickly," which glosses over the critical third part: they clicked a different result. Just returning to the SERP without a follow-up click is ambiguous — maybe the user got what they needed and is now browsing related results out of curiosity. It's the second click that resolves the ambiguity and confirms dissatisfaction.

"Pogo-sticking" entered SEO vocabulary in the late 2000s, borrowed from user-experience research that was studying how searchers navigate results pages. The pogo-stick metaphor captures the bouncing motion: down onto a result, up to the SERP, down onto another, up again.

Why Google cares

From Google's perspective, pogo-sticking is a high-quality user-quality signal because it's unambiguous. A searcher who clicks five results in sequence is telling Google, in behavior rather than words, that none of the first four satisfied the query. If the majority of searchers pogo-stick past a specific result before settling, that result probably doesn't deserve its current position.

The alternative: satisfaction by inference

Before pogo-sticking metrics, search engines had to infer satisfaction from indirect signals — dwell time, scroll depth, follow-on queries. All of these are noisy. Pogo-sticking is direct: the user explicitly rejected the result and picked another one. That's the clearest possible signal of dissatisfaction short of a thumbs-down button.

The catch

Pogo-sticking is only well-defined when the SERP has multiple clickable results to compare. Queries dominated by a single satisfying answer (brand-name navigational queries, for example) produce almost no pogo-stick data. The signal is most actionable on informational and commercial-investigation queries where searchers genuinely compare options.

Google detects pogo-sticking via SERP-side instrumentation. When a user performs a search, the SERP page itself loads a JavaScript bundle that logs events to Google's servers — clicks, back navigations, query entries, and returns to the SERP. The instrumentation is what allows Google to measure behavior they don't own — your destination page.

Click out

The initial click on a SERP result fires an event containing the query, the result URL, the position, the SERP features shown, and a session identifier. This is Google's anchor event for the pogo-stick measurement.

Return detection

Several mechanisms detect the return to the SERP:

• Browser back navigation that lands on a cached Google SERP URL — Google's JavaScript on the SERP fires on restore.
• Direct navigation back to google.com within a short window after the original click — session-attached.
• Retyping of the same or similar query within the session — Google's query-understanding model collapses near-duplicate queries into the same session.

If any of these occurs within the short-dwell window, the return is attributed to the specific result that was clicked.

Second click

If the user clicks a different result for the same query after returning, the second click is linked to the first in Google's session log, completing the pogo-stick pattern. Aggregated across thousands of sessions, this pattern becomes a statistically robust signal per (query, URL) pair.

Chrome telemetry (the wildcard)

There is an ongoing debate about whether Chrome browser telemetry feeds into pogo-stick detection in addition to SERP-side logs. Google has publicly denied using Chrome telemetry for ranking, but Chrome does have the ability to report navigation patterns. The conservative assumption is that SERP-side instrumentation does the heavy lifting and Chrome telemetry is used for aggregate field metrics (Core Web Vitals) rather than per-URL ranking signals.

Three related but distinct concepts. Conflating them causes most SEO miscommunication.

Bounce rate

A GA4 metric. Measures the percentage of sessions on your site that consisted of a single page view with no meaningful engagement. Calculated on your analytics, visible to you. Invisible to Google. A session that is a "bounce" in GA4 is often also a pogo-stick on Google's side, but not always — a user who read for 10 minutes and left is a bounce in GA4 but a long click on Google.

Dwell time

A Google-side metric. Measures the interval between SERP click and SERP return. Measured by Google's SERP instrumentation. Invisible to you. A low dwell time followed by a return to the SERP is the raw material of a pogo-stick; dwell time is the interval, pogo-sticking is the broader pattern that includes the subsequent click.

Pogo-sticking

The specific pattern of short dwell followed by a second click on a different result for the same query. Measured by Google. Invisible to you. Carries the strongest negative-signal weight of the three because of its unambiguous interpretation.

The inclusion hierarchy

Every pogo-stick is a short-dwell session. Not every short-dwell session is a pogo-stick (the user might have closed the tab without clicking a second result). Every pogo-stick session is a bounce in GA4. Not every bounce in GA4 is a pogo-stick (the user might have read for 8 minutes then closed).

Two Google systems use pogo-stick data directly. They do different things with it.

Navboost

Navboost is the ten-blue-links re-ranking layer. It takes Google's initial ranking (based on content, links, on-page, technical) and adjusts it based on accumulated per-query click behavior. Results with high pogo-stick rates get suppressed; results with high long-click rates get promoted. This is where most of the "why does that page outrank mine" magic actually happens on competitive queries.

Navboost operates on a rolling window — roughly 13 months in the leaked documentation — so pogo-stick signal from a year ago still counts, just with less weight than last week's.

Glue

Glue is Navboost's counterpart for non-ten-blue-links SERP features: featured snippets, People Also Ask boxes, image packs, video carousels. Pogo-stick data on SERP features adjusts whether Google shows those features at all for a given query, and which specific answer gets selected for the featured snippet.

If users consistently pogo-stick past a featured snippet, Google eventually demotes or removes it. If users who click a specific snippet don't return to the SERP, the snippet's source URL gets a quality boost.

The implication

Winning a featured snippet is valuable only if the snippet produces long clicks. A featured snippet that draws clicks but produces pogo-sticks is worse than no snippet — it exposes your URL to a large volume of dissatisfied-searcher signal. If you win a featured snippet and then see your overall ranking slide, the snippet is the likely culprit.

RankBrain is frequently misunderstood as "the thing that uses user behavior for ranking." It's not. RankBrain is a query-interpretation layer. It takes queries Google has never seen before (roughly 15% of daily queries) and maps them to related queries with known ranking profiles.

What RankBrain uses pogo-stick data for

Indirectly. When RankBrain maps a new query to a set of related known queries, the behavior data on those known queries feeds the ranking decision for the new query. Pogo-stick data is part of that, but it's filtered through the query-similarity mapping first.

What RankBrain does not do

RankBrain does not directly re-rank results based on pogo-stick signals in real time. That's Navboost's job. Confusion between the two is the most common conceptual error in SEO discussions of "algorithmic learning."

Why this distinction matters

If your ranking problem is on a well-established query with plenty of historical data, RankBrain isn't in the picture — Navboost is the relevant system. If your ranking problem is on a novel or long-tail query, RankBrain is mediating which similar queries' data gets applied, and the pogo-stick signals on those related queries matter.

Not directly. But you can estimate it with reasonable confidence using two proxies.

Proxy 1 — GSC position vs CTR pattern

If your URL ranks at position 4 for a query, your expected CTR based on position-curve averages is roughly 7-10%. If your actual CTR is 2-3%, users are seeing your snippet and not clicking — Navboost will eventually demote you based on the low click rate. If your CTR is 10% but your position has been sliding over the last 30 days, the clicks you're getting are producing pogo-sticks and Navboost is demoting you despite the strong CTR.

Proxy 2 — GA4 organic engagement time vs benchmarks

Pull engagement time for organic sessions on the specific URL. Compare to the URL's engagement time for non-organic traffic (direct, social, referral). If organic engagement is meaningfully shorter than non-organic, something about the SERP-to-page path is producing dissatisfied landings. Either the snippet is promising something the page doesn't deliver, or the page's first screen doesn't match what the snippet promised.

The combined diagnostic

Low CTR + declining position + shorter organic engagement than direct engagement = a near-certain pogo-stick problem. This combination correctly identifies pogo-stick issues in about 80% of cases in our consulting work.

Pogo-stick data is a two-way street. Most SEO work is defensive — improve your own URL's profile. Aggressive SEO also plays the offensive side.

Defensive: reduce your pogo-stick rate

Tighten the SERP-to-page match. Rewrite the first 100 words to immediately deliver what the snippet promised. Fix any first-screen issues (layout shift, interstitials, slow hero) that cause instant bounces. Match content shape to query intent. All of these reduce the rate at which real users pogo-stick away.

Offensive: increase the competitor's pogo-stick rate

This is what Bounce Rate Bot does. Engineered sessions search the target query, click the competitor, produce a short dwell, return to the SERP, click a different result. The same pattern Navboost uses to demote — generated synthetically at a volume large enough to move the aggregated signal.

The combined strategy

Defensive work takes 4-8 weeks to move rankings. Offensive work against a competitor takes 2-3 weeks. A combined campaign — fix your own signal while attacking the competitor's — compresses the timeline to 3-4 weeks for visible ranking gains on competitive queries, compared to 2-3 months with defense only.

Frequently asked questions about pogo-sticking and related signals.