What is a good engagement rate? (GA4 benchmarks)

GA4 counts a session as engaged when it meets any one of three conditions: it lasts longer than ten seconds of active attention, it triggers a key event, or it includes at least two page views. Engagement rate is engaged sessions divided by all sessions, and GA4's bounce rate is simply the same number flipped: a 58 percent engagement rate is a 42 percent bounce rate, one metric wearing two names. If you internalise only one thing, make it that the bar for "engaged" is low: ten attentive seconds or a second page.

Three sources publish engagement rate figures with their measurement described, and they agree more than benchmark sources usually do.

Databox's GA4 benchmark groups, medians from companies that connect their analytics accounts, put the overall median around 56 percent, with the fifteen industries it tracks spread between roughly 52 and 64 percent. AgencyAnalytics' platform data, drawn from more than 150,000 agency-managed campaigns as of December 2024, reports a median of 54 percent, and its companion bounce median of about 46 percent cross-checks arithmetically. A smaller sample from Orbit Media's site survey lands at 55 percent across 65 sites.

Three different SMB-leaning samples, one answer: the middle of the road is the mid-50s, and the well-performing band runs into the 60s. Figures much above 70 mostly belong to sites whose setup makes engagement easy to trigger, which is the next section. One agency table circulates with higher numbers, 63 to 71 percent, from 2020 client data with no client count disclosed; treat it as one shop's book of business, not a baseline. And the most-cited "study" on this SERP, an agency analysis claiming a 62 percent average, no longer exists at all: its URL redirects to a domain where the page 404s. This metric's benchmark literature is thin enough that the top citation is a dead link.

Engagement rate deserves less trust than its neighbours, because two configuration choices move it dramatically without any visitor behaving differently.

The timer is adjustable. The ten-second threshold that defines an engaged session can be raised in GA4's admin settings, up to sixty seconds. A property on the default timer and one set to sixty seconds can differ by tens of points on identical traffic, and no published benchmark states which setting its sample used, because sources aggregating connected accounts cannot know.

Key events move it too. Any session with a key event counts as engaged regardless of duration, so a site that marks a cheap event, a scroll, a video autoplay, as key inflates engagement mechanically. The metric partly measures your configuration taste.

Beyond configuration, honest structural differences do the rest: single-page content that answers in eight seconds earns an unengaged session while succeeding, multi-page stores clear the two-pageview bar by existing, and paid social traffic drags the rate down while returning readers pull it up. A 45 percent rate on a dictionary-style site can be better news than 65 on a store.

The same discipline as every benchmark on this site, with one addition specific to this metric. Compare against your own history, segmented by channel and device, and treat movement as the signal, an engagement drop on one channel usually means the traffic changed, not the site. Use GA4's own peer benchmarking, which includes engagement rate and compares your property against its real peer group's median and quartiles. And before comparing at all, check your own timer setting and key-event list, because a number you configured is only comparable to itself. For the fuller picture of what engagement is worth measuring, how to measure website engagement goes beyond the single rate, and the pre-GA4 sibling question lives at what is a good bounce rate.