Pages per session benchmarks: what the disclosed data says

Two publishers currently print pages-per-session figures alongside a real description of how they got them, fewer than for almost any other analytics metric. They measured different populations, so they do not merge.

E-commerce stores. Littledata's benchmark dataset puts the average at 2.6 pages per session across 2,800 stores benchmarked in 2023, with the top fifth of sites above 4 and the bottom fifth below 1.7. Shopify stores in the sample average 2.8. It is ecommerce-only and the data is a few years old, but the spread is the useful part: the distance between a struggling store and a strong one is bigger than the distance between any two industry averages.

Small and mid-sized businesses. Databox's benchmark groups report monthly medians from companies that connect their Google Analytics accounts: around 1.9 pageviews per session for B2B companies (over 800 contributing) and around 2.06 for B2C (roughly 500). Read those twice if the four-pages folklore had settled in: for a typical smaller site, two pages per session is not underperformance, it is the median.

One more source deserves mention for what it does not print. Contentsquare's 2026 Digital Experience Benchmark, the best-documented dataset in this space at 99 billion sessions across 6,500 mostly enterprise sites, reports the trend, page views per session down about 1 percent year over year, without publishing an absolute cross-industry figure on its free pages. When the most careful source declines to print the number every roundup confidently states, that tells you something about the roundups.

The 14-industry table that circulates, overall average 3.81, apparel above 9, descends from a single agency page that discloses no tool, no sample size, no time period and no metric definition. It is not necessarily wrong. It is unverifiable, which for a benchmark is the same problem, and its numbers sit far above what the disclosed sources report, which is exactly what you would expect if it drew on a handful of unusually engaged clients. When a table's provenance is unknown, the differences inside it, apparel versus finance, say, cannot be separated from noise.

The disclosed sources differ from each other too, and the reasons are instructive.

The metric is not one metric. Universal Analytics counted pages per session as page depth. GA4's views per session divides page and screen view events by sessions, repeat views of the same page included, and its enhanced measurement can count route changes in single-page apps as views, or not, depending on configuration. Two sites with identical behaviour can report meaningfully different counts on tagging choices alone.

The populations differ more. An enterprise retail sample, an SMB benchmark group and an ecommerce-only dataset describe different webs, and site type dominates everything: a blog that answers in one page earns a low count by succeeding, while a store's count rises with browsing. Contentsquare's data makes the point within a single industry: product detail pages alone drive around 40 percent of all page views on the retail sites it measures, and 61 percent of visits landing on one bounce. A number that swings that hard on page mix cannot carry a universal "good" threshold.

Your own history, first. Pages per session is most useful as a trend line, segmented by device and traffic source, where a drift down often means a landing page or navigation change did quiet damage, and a sudden rise can mean visitors are lost rather than engaged. Direction beats level.

Google's peer comparison, second. GA4's built-in benchmarking shows views per session against your actual peer group's median and quartiles, computed on the same metric definition, which no published table can promise.

And pair it with depth signals before celebrating movement. A rising page count with falling scroll depth is skimming, not engagement, and session duration moving the opposite way from page count is usually a navigation story worth investigating rather than a KPI to report.