Engaged sessions
Visits that stayed, went deeper or did something — not the ones that bounced.
What it means
Engaged sessions is the number of visits that lasted 10 seconds or more, fired a key event such as an add to cart, or reached two or more page views — any one of the three qualifies. It's a count of visits, not a rate. Everything else is a bounce, which makes its share of sessions and Bounce RateBounce RateA quick signal of landing-page relevance and speed.Under 25% is healthy the same measurement read from opposite ends.
Show the math
Formula and a worked example
How it's counted. Three ways to qualify: 10 seconds or more on site, a key event such as an add to cart or a purchase, or a second page view. A visit that does all three still counts once.
Worked example. 40,000 sessions in the month, 25,600 of them engaged — an engagement rate of 64%, so a Bounce Rate of 36%. Revenue of €80,000 is €2.00 per session, but €3.13 per engaged session. The second figure is what a visit is worth once it actually starts.
Watch the count and the rate together. Engagement holding at 64% while the count falls from 25,600 to 16,000 is a traffic collapse the percentage hides completely.
It answers the question
Of everyone who arrived, how many got past the door? A big count with weak revenue means people engaged and didn't buy; a small one means they never got that far.
Why it matters
It splits a bad revenue month in two. Few engaged sessions puts the problem before the page — the wrong visitors, a slow load, an ad that promised something the page doesn't show. Plenty of them puts it on the offer: price, delivery cost, stock, trust.
The 10-second bar is deliberately low, which is what makes it useful in one direction: a source that can't clear it is genuinely broken.
What good looks like
There's no benchmark for a count of visits — the right number depends on how much traffic your shop takes. Judge each row's engaged share against your site's own average and against that row's history. Rows sitting well below the site-wide share are the work list, and a count that falls while the share holds means you lost traffic, not quality.
How to improve it
| Lever | What you do | Expect | How long | Watch out for |
|---|---|---|---|---|
| Fast Fix the slowest landing pages | Compress hero images and cut render-blocking scripts on the three pages taking the most paid traffic | Engaged sessions up, Bounce Rate down on those pages | 1–2 weeks | Removing scripts can take reviews, personalisation or tracking tags with them — check what those pages lose before shipping it. |
| Fast Stop sending everyone to the homepage | Point each ad and each email at the page that matches what it promised | Engaged sessions up across the paid and email rows | 1 week | Every extra landing page has to be updated when prices or stock change, and stale ones outlive the campaign that needed them. |
| Slow Give a visit somewhere to go next | Add related products, size guides and reviews so there's a second thing to do after the first screen | More visits reaching two page views | 3–5 weeks | More links means more ways off the buying path — watch Conversion Rate on those templates, because a longer visit that ends without an order is worth nothing. |
| Slow Retire the sources that never engage | Cut placements and referrers whose engaged share sits far below your site average | A higher engaged share across the whole site | 1 quarter | Sessions and reach fall with them, and some of those cheap placements were doing top-of-funnel work you won't miss for months — and then will. |
Every lever costs something somewhere. The last column is the one to read twice.
Read it with
Engaged sessions is a count. Sessions is the count it came out of, and the pair is what tells you whether traffic or quality moved.
Growth that landed
More visits, and more of them real. The extra traffic behaved like the traffic you already had, which is what makes it worth scaling.
You bought arrivals
Extra visits that left inside 10 seconds. You're paying for impressions of your own homepage — the spend is real and nothing else is.
Cleaner, smaller
Less traffic and more of it engaged. This is what trimming broad placements looks like when it works.
Losing both
Fewer arrivals and a worse share of them staying. A broken tag or a site-wide speed problem produces this exact shape, so rule those out first.
Same measurement, opposite ends. The rate can sit still at 36% while the count halves, so the percentage says nothing changed and the count says half the business left. Only reading both catches it.
Breadth against depth. Lots of engaged sessions with a short average means people are clearing the 10-second bar and stopping there — a healthy count and a shallow visit at the same time.
Common misreads
Ten seconds, or a tab left open while someone answers the door, both qualify. Treat it as a filter that removes obvious junk, not as proof of intent.
They can't. Every session is one or the other. If they disagree, you're reading two different date ranges, two different rows, or a table filtered where the card isn't.
Source mix moves this without a word changing on the site. Add one broad prospecting campaign and the site-wide share falls while every existing page performs exactly as it did last month.
Also called
Engaged visits · non-bounce sessions · quality sessions
See yoursEngaged sessions per page, next to Sessions so you can read the engaged share for each landing page.
Open Google Analytics Pages →