Total impressions are the product of sessions and pages per session, which means session depth is a multiplier on everything else you do. A site with modest traffic and deep sessions can out-earn a larger site with shallow ones, and the arithmetic is simple enough to run on your own numbers in about five minutes.
The arithmetic
Ad revenue on a content site decomposes cleanly:
revenue = sessions
× pages per session
× viewable impressions per page
× revenue per viewable impression
Four terms, and publishers spend disproportionate effort on the first and last. Traffic acquisition and yield optimisation get the attention, while the two middle terms are treated as fixed properties of the site.
They are not fixed, and because the terms multiply, a proportional gain in any one of them is worth the same as the equivalent gain in any other. Moving pages per session from 1.4 to 1.7 has the same effect on total impressions as growing traffic by a fifth, and is usually cheaper to achieve.
Why the middle terms get neglected
Two reasons, both understandable.
Traffic is legible. Sessions are the number everyone reports, compares, and feels. Pages per session sits in an analytics screen most people glance at monthly.
Yield feels like the professional lever. Header bidding configuration, floor price management, and demand partner mix are where the specialist knowledge lives, so that is where attention goes. Meanwhile the reason a page earns half what it could may simply be that nobody continues past it.
Why this metric is a good predictor
Session depth is unusually informative because it is downstream of several things at once. A site with deep sessions almost necessarily has:
- Content that answers a question and creates the next one
- Pages fast enough that a second load is not a deterrent
- An ad experience tolerable enough that readers continue
- Internal linking that puts the next step where the reader finishes
- Traffic arriving on browse-type queries rather than purely reference ones
Each of those independently affects revenue. Session depth moves when they move, which is what makes it a useful single indicator rather than just another engagement statistic.
What it does not tell you
Being careful here, because a metric treated as a target stops being a good metric.
- It says nothing about traffic quality by geography. Deep sessions from markets with low advertiser demand may be worth less than shallow sessions from high-demand ones.
- It can be inflated destructively. Paginating one article across five screens raises the figure and worsens the site. Any tactic that increases depth by making the reader work harder is extracting from the same reader rather than serving them better.
- Comparisons across categories are meaningless. As covered in session depth by content vertical, structure dominates, so benchmarking against a different category tells you nothing.
- It is not the whole picture for return-driven properties. A news site with shallow sessions and frequent returns can be perfectly healthy.
How to work out what it is worth to you
The useful exercise, which takes about five minutes and is specific to your site rather than to any benchmark.
- Take your current pages per session for a single property, over a stable recent period.
- Take your revenue for that same period and divide by sessions to get revenue per session.
- Model a modest improvement, say a tenth of a page. Multiply through and see what that is worth annually.
- Compare that figure against the cost of the equivalent traffic increase. For most content sites the depth improvement is considerably cheaper, because it requires editorial and structural work rather than acquisition spend.
That comparison is the whole argument. Not that session depth is more important than traffic, but that it is usually the cheaper unit of improvement and it is the one nobody is working on.
Where the improvement usually comes from
In descending order of reliability:
- A real next step at the point of completion. The specific question a reader has after finishing, answered by a specific link, placed where they finish. This outperforms every generic recommendation module.
- Faster second page loads. Patience is lower on the second page than the first.
- Removing whatever is worst about the ad experience. Usually one particular unit or behaviour rather than the overall count.
- Content clusters rather than isolated articles. A group of pages covering a topic properly gives the reader somewhere to go that is genuinely relevant.
- Targeting browse-intent queries alongside reference ones, which changes the mix of visitors arriving.
None of that is exotic. It is mostly editorial discipline, and it compounds with every other improvement because of where it sits in the arithmetic.
Frequently asked questions
Why does pages per session matter for ad revenue?
Because total impressions are the product of sessions and pages per session. Session depth acts as a multiplier on everything else, so a proportional gain in depth affects total impressions the same way an equivalent gain in traffic does, and is usually cheaper to achieve.
Is it better to increase traffic or session depth?
It depends on the relative cost, which is specific to your site. For many content publishers, improving session depth through editorial structure and internal linking is cheaper than acquiring the equivalent additional traffic, and it compounds with other improvements because of where it sits in the revenue arithmetic.
Can pages per session be a misleading metric?
Yes. It can be inflated by paginating single articles or otherwise making readers work harder, which raises the number while worsening the site. It also ignores differences in advertiser demand by geography and is not comparable between content categories.
What actually increases pages per session?
Offering a genuine next step that answers the question the reader has after finishing, placed where they finish reading. Faster second-page loads, a less obtrusive ad experience, and building topic clusters rather than isolated articles also contribute reliably.
Should a news site worry about low session depth?
Usually not. News properties typically produce shallow sessions with frequent returns, so visit frequency rather than depth is the appropriate measure. Judging that model by pages per session measures the wrong thing.
How do I calculate what a session depth improvement is worth?
Take pages per session and revenue per session for one property over a stable period, model a modest increase such as a tenth of a page, multiply through to an annual figure, then compare that against what the equivalent traffic increase would cost to acquire.
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