What is a good bounce rate for product pages?

Bounce rate used to mean a session with a single page view and no other interaction. That definition made a page look bad even when the visitor read it, got what they needed and left satisfied.

In current analytics the definition is different. A session counts as engaged if it lasts beyond a short threshold, or produces a key event, or includes more than one page view. Bounce rate is the share of sessions that were not engaged, which makes it the exact inverse of engagement rate.

Two consequences follow. First, any target number quoted from the older definition is not comparable to what your reports show now. Second, the metric can be moved without improving anything, for example by making a page load in a way that keeps a session technically alive.

That does not make it useless. A page where almost nobody engages at all is worth looking at. It makes it a rough signal rather than a target, and it means chasing a specific figure is chasing an artefact of the definition.

So do not set a bounce rate goal. Set goals on the behaviours that produce orders.

The question a product page answers is whether a visitor moves towards buying. So measure the movement.

Add-to-cart rate per product view is the primary number. It counts only people who reached a product page, so everyone in the denominator saw the thing you are asking them to add.

Next, the mix of next pages. Where do people go from a product page. To the cart is good. To another product is usually fine, it means they are comparing within your range. To a policy page can mean shipping or returns are unclear. Back to a collection page repeatedly can mean the page did not answer the question and they are searching for a different option.

Then scroll depth and time on the key sections, which tell you whether the answers you added are being reached at all.

And revenue per product view, which combines the page's ability to convert with the value of what it sells.

Those four tell you what to change. Bounce rate tells you only that something happened or did not, without saying what.

There are three situations where the number is worth acting on, and they all involve comparison rather than an absolute value.

A single page much worse than similar pages. If one product page shows far less engagement than the others in the same category, something specific is wrong with it: a broken image, a missing price, an out of stock variant selected by default, a script failing.

A single traffic source much worse than the others. Paid traffic landing on a product page with very low engagement usually means a mismatch between the ad and the page, or bot traffic, or a landing page that is not what the campaign promised.

A sudden change on the same pages. Engagement falling across a template overnight almost always means a release broke something. Check the date against your deployment log before looking for a marketing explanation.

Outside those, a difference of a few points between pages is noise. Also remember that a page answering a question well can produce a short, unengaged session and a happy visitor, particularly for stock checks, delivery information and returns lookups by existing customers.

A surprising share of engagement problems are recording problems, and they are quick to check.

Duplicate analytics installations inflate page views and can make every session look engaged, which hides real problems. One installation, verified.

Consent banners change what is recorded and when. If a banner blocks measurement until accepted, the sessions of people who ignore it may be recorded partially or not at all.

Redirects and in-app browsers can start a new session mid journey, splitting one visit into two and making both look worse.

Bot traffic inflates the denominator with visits that were never going to engage. Check the source report for sudden spikes from unusual referrers.

And check the engagement threshold configuration if your setup allows it to be changed, since altering it changes every historical comparison.

Once those are ruled out, compare like with like: the same template, the same traffic source, the same device, before and after a change. That comparison is meaningful even when the absolute number is not.

The fixes are the same ones that lift add-to-cart rate, because that is the real problem underneath.

Speed on a phone over mobile data. If the page takes several seconds to become usable, a share of visitors never see it at all, and no amount of copy will help them.

The first screen. Price, variant options, delivery cost and window, and the add-to-cart button should be reachable without three swipes.

The unanswered question. Sizing, materials, compatibility, what is in the box, how long delivery takes, what happens if it is wrong. Every one that is missing sends someone away to find out elsewhere.

The mismatch. If paid traffic bounces and organic does not, the ad promised something the page does not show quickly. Fix the landing page or the ad, not the theme.

Evidence. Reviews placed next to the relevant decision, not only in a block at the bottom.

Make one change at a time on a group of similar pages, and measure add-to-cart rate per product view before and after. That number will tell you whether it worked. The bounce rate will follow it anyway.