Active users
The people who did something on a row, not the ones who only loaded it.
What it means
Active users is the count of distinct people who engaged, counted once per row for the period. Engaging means a visit lasting 10 seconds or more, a key event, or a second page view — anyone who lands and leaves in three seconds is excluded. It's people rather than visits, split by page.
Show the math
Formula and a worked example
How it's counted. One count per person per period, on each row where they had at least one visit that counts as an engaged sessionEngaged sessionsEngaged sessions: sessions lasting 10s+, with a key event, or 2+ page views. Coming back five times on the same browser still counts once.
Worked example. The Pages table shows 22,000 active users on one collection page and 8,000 on another. The first brought €60,000, the second €48,000. That's €2.73 per active user on the busy page against €6.00 on the quiet one — most of the people, less than half the value each.
Read the column that way and the rows with plenty of active users and little revenue stand out. That's where attention is going and money isn't.
It answers the question
How many real people engaged with this page? A row can carry heavy traffic and few active users, and that gap is arrivals that never engaged.
Why it matters
Volume counts flatter bad sources. A bot-heavy referrer or a misfiring redirect inflates visits and page views without a single person reading anything, while Active users only counts people who stayed long enough to do something. That makes it the safer column when you compare one page against another.
It's also the right denominator for value per person. Revenue divided by active users tells you what an engaged person is worth on that page, and that figure decides where the next budget goes far better than raw traffic does.
What good looks like
No target exists for a count — how many active users is good depends on the size of your shop. Compare rows against each other on the same table and each row against its own last few months, then read it beside what that row earned. A big active-user count next to small revenue is a row to work on, not a row to celebrate.
How to improve it
| Lever | What you do | Expect | How long | Watch out for |
|---|---|---|---|---|
| Fast Hold the popup back | Delay the newsletter overlay until after the first scroll or 15 seconds | Fewer instant exits, active users up on landing pages | Days | Sign-ups fall with it, so the flows fed by that list have a smaller audience to earn from next quarter. |
| Fast Make the first screen appear | Compress hero images and defer render-blocking scripts on the templates taking the most traffic | More arrivals clear the engagement bar | 1–2 weeks | Deferring scripts can silence review widgets, personalisation and tags — confirm what still fires before you judge the lift. |
| Slow Match the landing page to the promise | Send each campaign to a page that repeats its offer instead of the homepage | Active users up on those landing pages | 3–5 weeks | Every extra landing page needs maintaining when prices or stock change, and the upkeep lands on whoever runs the site. |
| Slow Retire the rows that never engage | Stop sending paid traffic to pages showing visits and almost no active users | A higher engaged share across the site, less wasted spend | 1 quarter | Total sessions and reach both fall, which looks bad on a traffic chart and can quietly shrink the retargeting pool you fill from. |
Every lever costs something somewhere. The last column is the one to read twice.
Read it with
Active users tells you how many engaged. It takes a second column to say whether that engagement was worth anything.
Attention that pays
Real people engaging and buying. The row to grow, as long as the cost per person holds when you scale it.
Busy and broke
They engaged and didn't buy. That points at price, stock or a mismatch between what the ad promised and what the page shows — not at the volume.
Fewer people, better ones
A small, keen audience carrying real money. Often a narrow placement or a warm list — good while it lasts, hard to grow without diluting.
Going quiet
The row is fading at both ends. Usually a campaign that stopped, a feed that broke, or a tag that no longer fires on those landing pages.
The distance between them is the crowd that arrived and did nothing measurable. 30,000 total against 22,000 active means roughly a quarter left before anything happened — a drop-off neither number shows on its own, and the cheapest one to fix.
Breadth against depth. Plenty of active users with a short average means lots of people clearing a low bar and going no further, which reads as engagement in one column and as a bounce problem in the other.
Common misreads
They never will. Everyone who arrived and left inside a few seconds counts in one and not the other, and that gap is the point of having both.
People are counted once per row and once overall. Someone who engaged with two pages appears twice down the column and once in the site figure, so the rows always sum higher.
The bar can move instead. A slower page pushes visits under 10 seconds and a key event that stops firing removes another route to qualifying — arrivals never changed, only what counted as engagement.
Also called
Users · engaged users · active visitors
See yoursActive users per page, next to what each page earned, so you can see which pages hold people who do something.
Open Google Analytics Overview →