Session depth varies enormously between content categories, and almost none of that variation is about content quality. It is determined by whether the reader's question is answered completely by one page, and whether there is an obvious next question. Recipe and reference sites behave differently from news for structural reasons, and no amount of related-posts optimisation overrides that.
The variable that actually decides it
Publishers comparing their pages per session against some industry benchmark are usually comparing themselves against a different structural situation and drawing the wrong conclusion.
The question that predicts session depth is not "is this content good". It is: when the reader finishes this page, do they have another question?
Everything else follows from that.
Why the categories differ
Single-answer content
Currency conversion, a specific statistic, a definition, today's price of something. The reader had one question, the page answered it, and they leave satisfied.
These pages produce shallow sessions by design, and that is not a failure. Trying to force depth here usually means obstructing the answer the reader came for, which damages the thing that brought them in the first place.
Comparison and selection content
The opposite situation. Someone choosing between options genuinely has a next question, and then another. This category naturally produces deeper sessions because the reader's task is not complete after one page.
Planning content
Travel, events, projects, anything where the reader is assembling something. A person planning a trip needs where to stay, what to do, how to get around, and when to go. Each answer generates the next, which is why this category tends to be among the deepest.
Recipe and instructional content
Interesting because it splits. Someone looking for one specific recipe is single-answer. Someone browsing for ideas is in selection mode and goes deep. The same site serves both, and the ratio between them is largely determined by which search queries you rank for rather than anything on the page.
News and time-sensitive content
Typically shallow per visit, with the compensating characteristic that readers return frequently. The value is in visit frequency rather than session depth, which means judging a news property by pages per session measures the wrong thing entirely.
The search intent problem
Here is the part that catches publishers out.
Two sites in the same nominal category can have very different session depth because they rank for different types of query. A food site ranking mainly for "how many grams in a cup" will look shallow. A food site ranking for "easy weeknight dinners" will look deep. Both are food sites.
So before concluding that your engagement is poor, look at what queries actually bring people in. You may be comparing a reference-query site against a browse-query site and blaming your layout.
This connects directly to how AI answers are changing search. Single-answer queries are precisely the ones most likely to be resolved on the results page without a click, which means the shallow-session traffic is also the traffic most exposed to that shift.
What actually changes session depth
In rough order of effect:
- Which queries you target. By far the largest lever, and the one publishers treat as fixed. Pursuing browse and comparison queries changes session depth structurally in a way no design change can match.
- Whether a genuine next step exists. Not a related-posts grid, but an actual answer to the question the reader now has. The difference between "you may also like" and "next, work out what size you need" is enormous.
- Whether the next step is visible where the reader finishes. Links buried in a sidebar the reader never reaches do nothing.
- Page speed. The second page load is where impatience shows up. A reader who tolerated a slow first page frequently will not tolerate a slow second one.
- Ad experience. Covered in ad density and page views, and a real factor in whether anyone continues.
When shallow is correct
Worth saying plainly, because the industry treats session depth as a virtue in itself.
If your audience arrives with a specific question and leaves with the answer, a shallow session means you did your job. Interfering with that to manufacture depth, by paginating a single answer across several screens or withholding part of it, damages exactly the reputation that produced the traffic.
The right response to a structurally shallow property is not to fight the structure. It is to monetise appropriately for it and to build return visits instead, which is a different and often more durable asset.
How to look at this on your own sites
- Segment pages per session by landing page type, not site-wide. Reference pages and browse pages should be assessed separately.
- Cross-reference with the queries each page ranks for. That usually explains the difference immediately.
- Compare like against like across your properties. A travel property and a data property are not comparable, and averaging them produces a number that describes nothing.
- Track returning visitor rate alongside depth. Shallow and frequent can be worth more than deep and occasional.
Frequently asked questions
Why do some content sites have much deeper sessions than others?
Mainly because of whether the reader's question is fully answered by one page. Single-answer content such as a definition or a specific figure produces shallow sessions by design, while planning, comparison, and selection content generates a natural next question and therefore deeper sessions.
Is a low pages-per-session figure a problem?
Not necessarily. If visitors arrive with one specific question and leave with the answer, a shallow session means the page worked. Artificially manufacturing depth by splitting answers across pages damages the reputation that generated the traffic in the first place.
What is the biggest lever on session depth?
Which search queries bring people in. Two sites in the same category can differ substantially in session depth purely because one ranks for reference-style queries and the other for browse-style queries. This matters far more than layout or related-content modules.
Do related posts modules increase pages per session?
Modestly at best. What works better is offering a genuine next step that answers the question the reader now has, positioned where they finish reading, rather than a generic grid of loosely related articles placed in a sidebar.
How should session depth be measured across a portfolio of sites?
Segmented by landing page type and cross-referenced against the queries each page ranks for. Averaging across properties in different categories produces a figure that describes none of them and supports no decision.
Are shallow sessions more exposed to AI search changes?
In general yes, because single-answer queries are the ones most readily resolved directly on the results page without a click. Content whose value lies in one quick fact faces more pressure than content that supports a longer decision process.
West Coast Media Solutions Inc. operates a network of owner-operated content websites and builds sites for clients across Canada. Standards cited are published industry definitions; the patterns described reflect general practice rather than figures from any particular property, and your own reporting is the only measure that matters for your inventory.
Interested in the network?
We run our own portfolio of niche content sites and sell display and native inventory across them.
See the Network