Definition
For a single page view, scroll depth is the furthest point the visitor brought into view, divided by the length of the page. A visitor who scrolled until the bottom edge of the screen reached the midpoint of the page has a scroll depth of one half for that view.
Furthest point scrolled ÷ Total page length × 100(%)
- Furthest point scrolled is the lowest position on the page that appeared on screen during the view, measured from the top. It includes the first screen, so a visitor who never scrolls still has a depth equal to the share of the page visible on arrival.
- Total page length is the full height of the page as rendered on that device. The same page is much longer on a phone than on a wide screen, and pages that load more content as the visitor scrolls change length while they are being measured, so the figure depends on how much of the page had loaded when it was taken.
Threshold reporting
Few tools store the exact depth of every view. Instead they fire an event when a view crosses a set point and report the share of page views that reached it. In GA4Google's event-based analytics platform., the scroll event collected by enhanced measurement fires once per page when the visitor nears the bottom, which answers only one question: did they get close to the end? Tag managers offer a scroll depth trigger that fires at whatever percentage or pixel thresholds the team chooses, and the guide to tags, triggers and variables shows how it can be combined with a timer so that only a view that both scrolled far and stayed counts as engaged. Before trusting such events, the guide to debugging with preview mode shows how to watch each threshold fire as you scroll a test page.
The result is a short distribution per page: nearly every view reaches the first threshold, fewer reach each later one, and the drops show where attention runs out.
How to interpret it
Scroll depth has no direction that is good everywhere. Deeper is better on a long article or a page whose call to action sits low; it is irrelevant on a page whose purpose is met on the first screen, such as a contact page with the phone number at the top. Useful readings:
- By template. Articles, product pages and Landing PageThe page a visitor arrives on from a campaign, built to turn that visit into one action. designs differ in length and job, so a site-wide average mixes unlike things.
- By page length. Depth falls as pages get longer, so a long guide and a short news item cannot be compared directly. Compare pages of similar length, or look at whether a specific element was reached.
- By source and device. Visitors from search who found the answer near the top behave differently from newsletter readers who came to read the whole piece, and phones produce longer pages and more scrolling.
- Against the position of what matters. The most useful question is often not how deep the average view went but whether the threshold just above the form, price or booking section was crossed.
The guide to measuring content performance places scroll depth with engagement measures rather than outcomes and warns that it reflects layout as much as interest: a diagnosis of where a page loses people, to be confirmed by the outcome.
A walking holiday company moves its booking section
Stilecross Walking Holidays published 22 route pages, each about 3,000 words long with dates and prices in a booking section at the very bottom. Route pages drew 9,400 sessions a month but only 31 enquiries. The team added scroll thresholds at 25%, 50%, 75% and 90%: 88% of page views crossed 25%, 46% crossed 50% and 14% crossed 75%, while the booking section began at 92% of the page on phones. They moved a short summary of dates, prices and an enquiry button to just below the route introduction, at about 20% of the page, and left the detail below. Over the next two months route-page sessions were unchanged and enquiries rose to 58 a month. Scroll depth itself fell slightly, because fewer visitors needed to reach the bottom.
What moves it
- Page length. Longer pages produce lower average depth with no change in interest.
- Layout and the first screen. A large banner with nothing to read, or a design that looks finished above the fold, stops visitors scrolling; clear signals that more follows encourage it.
- Intent and match to the query. Visitors who find their answer near the top have no reason to scroll, and visitors who arrive on the wrong page leave from the first screen.
- Load speed and stability. Slow or shifting pages lose visitors before they start scrolling.
- Measurement setup. Thresholds, lazy-loaded content that changes page length and duplicate triggers all change the reported figure without any change in behaviour.
What good looks like
There is no reliable benchmark for scroll depth, and the site does not cite one. The figure depends so heavily on page length, layout, device and why visitors arrived that a number from another site or another template says nothing useful about this one. Judge it internally: against the same template's history, against pages of similar length on the same site, and above all against where the page's important elements sit. If the threshold just above the call to action is rarely crossed and the page is meant to produce enquiries, the depth is too low for that page's job, whatever the average says. If enquiries are healthy, a modest scroll depth is no cause for concern.
Related metrics
Engagement RateShare of sessions that were engaged; on social, interactions relative to reach. counts the share of sessions that met GA4's engagement conditions, a broader signal that scroll depth can help explain. A SessionA group of one user's interactions with a site or app in one visit, ended by inactivity. may contain several page views with different depths, so scroll figures are usually read per page view rather than per session. In GA4Google's event-based analytics platform., enhanced measurement records only the near-bottom threshold; finer thresholds need a tag manager. On a Landing PageThe page a visitor arrives on from a campaign, built to turn that visit into one action., scroll depth shows whether visitors reached the offer at all, which helps separate a page that fails to persuade from one that is never read.
Where scroll depth belongs in a reporting plan depends on the objective. The website traffic objective suggests deciding in advance what an engaged visit must include, and scroll depth is one of the options. In the See-Think-Do-Care framework it is one of the attention measures suited to content aimed at people not yet ready to buy. It should not be promoted to an outcome: the guide to key events and conversions in GA4 lists marking scroll depth as a key event among the common mistakes, because it inflates conversion counts and weakens any bidding signal built on them.
Common mistakes
- Treating depth as reading. A visitor can scroll to the bottom in seconds; depth shows exposure, not attention.
- Comparing pages of different length. A long guide will always look worse than a short post.
- Marking a scroll threshold as a conversion. It counts behaviour, not outcomes, and pollutes bidding data.
- Assuming the default event covers it. GA4's built-in scroll event reports one threshold near the bottom only.
- Ignoring device. Phone and desktop depths describe different page lengths and should be read separately.
- Chasing a higher number. Moving important content up can lower depth and raise results, which is the better trade.
