Session depth is a metric most retailers track and most retailers misread.
The standard framing is: high pages-per-session is good, low pages-per-session is bad. If people are visiting more pages, they're more engaged. If they're visiting fewer pages, something is wrong.
That's not wrong, but it's incomplete in a way that leads to incorrect interventions. The question isn't "how many pages did this shopper visit." It's "why did they visit that many pages, and did those pages serve them."
Two Kinds of Session Depth
When we look at session data from retail stores, there are two distinct patterns that both produce high pages-per-session numbers, and they have opposite implications for conversion.
The first pattern is exploratory depth. A shopper lands on a category page, finds a product that matches their intent on the first scroll, clicks through to the product detail page, adds to cart, and completes the purchase. The session might be 4-6 pages total. Pages-per-session is moderate. Conversion rate is high. Time-on-site is efficient.
The second pattern is search depth. A shopper lands on a category page, doesn't find what they're looking for in the first scroll, keeps paginating or filtering, visits 8-12 product detail pages looking for something that fits, and either eventually purchases or exits without buying. Pages-per-session is high. Conversion rate is low. Session duration is high because the shopper is working hard to find something.
If you optimize purely for pages-per-session as a positive metric, you could accidentally build a store experience that generates a lot of search depth without generating exploratory depth. More page views, more time on site, lower conversion.
Session Depth as a Leading Indicator
The more useful way to think about session depth is as a leading indicator of product-to-shopper fit.
When the first page a shopper lands on contains products that are genuinely relevant to their taste and intent, something specific happens: they don't need to work hard. They find items worth clicking on in the first few positions. The product detail pages they visit are ones they're actually considering. Add-to-cart happens earlier in the session.
When this pattern is consistent, pages-per-session tends to be in a moderate range, dwell time per page is higher on the product detail pages, and conversion rates are strong. The ratio of product detail page visits to add-to-cart events is efficient. You can see this in the data as a low "views-to-add" ratio.
When product-to-shopper fit is poor, you see the inverse: the shopper is paging through category results looking for something that matches. High category page pagination, high product detail page visit count, low add-to-cart relative to PDP views. The session is long but it's effortful exploration, not confident purchase progression.
Where This Shows Up in Category Analytics
We look at a specific metric when evaluating how a category page is performing for different shopper segments: the "depth-to-purchase" ratio. This is pages visited in a session on the way to a conversion, divided by pages visited in sessions that end without a conversion.
In a well-personalized category, converting sessions and non-converting sessions have similar depth profiles. A shopper who buys visited 4 pages. A shopper who didn't buy also visited 4 pages. The difference is in what happened on those pages, not how many there were.
In a poorly matched category, converting sessions are shallower and non-converting sessions are much deeper. The shopper who bought found something quickly. The shopper who left was searching for a long time and gave up.
That gap between converting and non-converting session depth is a signal that a large portion of your sessions are in search mode, not exploration mode. It's a category relevance problem, not a UX problem.
Improving session depth metrics by adding more pagination pages, more filter options, or a better search bar addresses the symptom, not the cause. If the products shown on the first page aren't relevant to a meaningful portion of your visitors, more ways to search through the irrelevant products doesn't fix the underlying mismatch.
How Personalization Changes the Depth Pattern
When category page sort order is personalized based on purchase history, the products shown in the first scroll are the ones this specific shopper is most likely to buy. The practical effect on session depth is counterintuitive: average pages-per-session often goes down, not up, in the first few weeks after personalization is enabled.
This is because a significant share of high-page-count sessions were search sessions. They were long because the product match was poor. Once the first page shows relevant products, those sessions get shorter. The shopper finds what they need faster.
But conversion rate goes up at the same time, and this is the critical thing to watch. You might see a dashboard where pages-per-session drops and conversion rate rises simultaneously. Without the context of "two kinds of session depth," this looks confusing. It's actually a signal that things are working correctly. Search-depth sessions are being replaced by exploratory-depth sessions.
The ratio to watch is not pages-per-session in isolation. It's pages-per-session relative to add-to-cart rate. When personalization is working, both move in the direction of efficiency: fewer pages needed to reach an add-to-cart, and more sessions reaching add-to-cart at all.
Applying This to Your Category Analytics
If you want to audit your own category pages using this framework, the starting point is a segment comparison.
Pull your category analytics split by new and returning visitors. Compute the average pages-per-session and the conversion rate for each segment on your highest-traffic category pages. In most stores we've looked at, returning visitors have higher pages-per-session and lower conversion rates on category pages than first-time visitors. This is the signature of a search-depth problem: your most loyal shoppers are having to search hardest to find what they want, because the category sort isn't built around them.
First-time visitors often convert better from category pages despite having lower session depth, because the bestsellers-by-volume sort that most stores default to is actually well-matched to first-time visitor behavior. First-timers are exploring the category for the first time. The bestsellers list tells them what's popular, which is a reasonable first purchase signal when you have no other information about this person.
Returning visitors are a different case. They already know what they like. They're coming back to find more of it, or to see what's new that fits their established taste profile. Showing them the same bestsellers list they saw on their first visit is a missed opportunity.
The Metric That Matters Most
If we had to pick one leading indicator to watch when evaluating category page personalization quality, it would be the conversion rate difference between returning and new visitors on category pages. When that gap is large (returning visitors converting at a much lower rate), the category is performing for the median shopper and failing the loyal one. When personalization closes that gap, it's working.
Session depth is a useful secondary signal, particularly the depth-to-purchase ratio described above. But conversion rate disparity between visitor segments is the clearest diagnostic of whether your category pages are serving the people who know your store best.