Dispatch
What Is Dwell Time? The Engagement Signal Nobody Can Fully Measure
Dwell time is the amount of time a person spends on a page after clicking through from a search results page, before returning to those results. That return trip is...
On this page
- How dwell time differs from metrics you can actually see
- What Google has actually said
- Bing’s position is different, and more publicly engagement-focused
- Why dwell time still gets discussed so much despite the uncertainty
- What you can actually do, given the measurement gap
- The practical takeaway
- Related posts:
Dwell time is the amount of time a person spends on a page after clicking through from a search results page, before returning to those results. That return trip is the key part of the definition. It’s not just “time on page.” It’s specifically the time between leaving the search results and coming back to them, which means it only exists in relation to a search engine’s results page in the first place, not to any visit that didn’t start there.
That narrow definition is also exactly why dwell time is so hard to pin down with confidence: search engines control the click and the return, not you. Your own analytics can tell you how long someone stayed on a page. It generally cannot tell you whether they then went back to a Google or Bing results page versus closing the tab, opening a new one, or navigating elsewhere entirely. The actual measurement happens on the search engine’s side, where you have no visibility.
How dwell time differs from metrics you can actually see
It’s worth being precise about this because the terms get used loosely:
| Metric | What it measures | Where it's measured |
|---|---|---|
| Dwell time | Time between a search-result click and a return to that results page | On the search engine's side; not directly visible to site owners |
| Time on page | Time spent on a single page, regardless of where the visitor came from | Your own analytics (with caveats about how exits are tracked) |
| Session duration | Total time across all pages in a session | Your own analytics |
| Bounce rate (GA4) | Inverse of engagement rate; depends on a 10-second threshold, key events, or 2+ pageviews | Your own analytics |
| Engagement rate (GA4) | Percentage of sessions meeting GA4's engagement criteria | Your own analytics |
Time on page and session duration are the closest available proxies for dwell time that you can actually access through your own tools, but they are not the same measurement, because they don’t account for where the traffic came from or whether the visitor returned to a specific search results page afterward.
What Google has actually said
Google has not confirmed dwell time as a ranking factor, and on at least one occasion has pushed back on the idea directly. In a Reddit AMA, Google’s Gary Illyes dismissed theories that dwell time and click-through rate function as ranking signals, reportedly stating, “Dwell time, CTR, whatever Fishkin’s new theory is, those are generally made up crap,” adding that “search is much more simple than people think.” That’s about as direct a denial as a Google representative has given on this specific topic, and it’s worth taking at face value rather than assuming Google is just being cagey.
That denial doesn’t mean user behavior plays no role at all in how Google evaluates search quality broadly. Google has separately acknowledged using click data and related signals for evaluation and quality-assessment purposes, distinct from a direct ranking calculation. The honest summary is: Google says it doesn’t use dwell time as a ranking signal, and there’s no verified evidence contradicting that statement, even though it remains a popular theory in SEO circles, partly because correlation between engagement and rankings is easy to observe and easy to mistake for causation.
Bing’s position is different, and more publicly engagement-focused
Bing has been more transparent than Google about treating engagement signals, including dwell time, as something it actively watches. In 2011, Bing’s then-webmaster outreach lead Duane Forrester published a post on the Bing Webmaster Blog explicitly framing dwell time as a quality indicator Bing pays attention to, the idea being that a visitor who lands on a page and sticks around is a reasonable proxy for that page actually answering their need, while a visitor who clicks back almost immediately suggests the opposite. That post is widely cited in SEO industry coverage, including by Search Engine Journal’s analysis of dwell time as a ranking factor, as the origin of “dwell time” as a term in SEO discussion.
It’s worth being careful about exactly what this establishes. Bing describing dwell time as “a signal we watch” is not the same as Bing publishing a precise weighting formula or guaranteeing it functions as a direct, isolated ranking input. The honest position is narrower than “Bing confirmed dwell time as a ranking factor” and closer to “Bing has been more open than Google about treating time-based engagement as one input into how it assesses content quality.” Treat any claim more specific than that, including round numbers like “Bing wants 90+ seconds per page,” with skepticism unless you can trace it back to a specific, current Bing statement, since recycled SEO-blog folklore tends to harden into specific-sounding numbers over time without actually getting more accurate.
Why dwell time still gets discussed so much despite the uncertainty
Even without confirmation that it’s used directly, dwell time (or its closest proxies) tends to correlate with things that obviously matter for other reasons: content that actually answers the query, pages that load quickly enough not to lose impatient visitors, and layouts that don’t frustrate readers into leaving early. Optimizing for those things is good practice regardless of whether a search engine is counting the seconds, which is part of why “optimize for dwell time” persists as advice even in the absence of a confirmed mechanism: a lot of the underlying actions are sound on their own merits.
That’s a different claim from “dwell time is secretly a ranking factor Google won’t admit to,” and it’s worth keeping the two separate. The case for working on these things doesn’t need a confirmed ranking mechanism behind it.
What you can actually do, given the measurement gap
Since you can’t measure true dwell time directly, the practical move is to focus on what’s known to influence the proxies you can measure (time on page, scroll depth, return visits) and to assume those correlate reasonably well with whatever search engines might be observing on their end:
- Match content depth to query intent. A quick-answer query deserves a quick, clear answer near the top of the page, not three paragraphs of preamble before the actual information. A genuinely complex topic deserves real depth, not padding.
- Front-load the value. Visitors who don’t find what they’re looking for within the first screen or two tend to leave, regardless of how good the content further down the page is.
- Keep load times fast. A page that’s slow to become usable loses visitors before they’ve had any chance to engage with the content at all, which drags down every downstream engagement metric you might otherwise improve.
- Use format deliberately. Long-form text, video, interactive tools, and short reference answers all serve different intents; matching format to what the query actually calls for tends to keep visitors engaged longer than defaulting to one format for everything.
- Remove friction. Intrusive interstitials, aggressive pop-ups, and disruptive ad placements push visitors to leave faster, working directly against whatever engagement signal, confirmed or not, you’re trying to support.
The practical takeaway
Dwell time is a real, well-defined concept, time between a search click and a return to the results, but it’s one that site owners can’t directly measure and that Google has explicitly denied using as a ranking signal. Bing has been more open about watching engagement-type signals, dwell time included, though even that is closer to “we pay attention to this” than a precisely documented ranking formula. The most defensible approach is to treat dwell time as a useful mental model for what good content tends to produce (sustained, genuine engagement) rather than as a lever you can pull directly, and to optimize for the underlying causes, relevance, depth, speed, format, and friction, that would matter to readers and to search quality whether or not any specific algorithm is counting seconds.