Sessions (GA4)
How many visits your website took, before you judge anything measured as a share of them.
What it means
Sessions (GA4) is the number of visits to your own website in the period, as counted by Google Analytics. It counts visits, not people — Total UsersTotal UsersCompare with sessions to gauge how often people return. is the separate card for distinct visitors, so one person coming back three times adds three here and one there. The name carries its source for a reason: Amazon listing visits are a different number, on a different site, under a similar name.
Show the math
Formula and a worked example
A visit is one arrival on your website, counted by Google Analytics rather than by Shopify or by this app. Each source counts with its own rules, which is why the three never agree exactly.
Worked example. March takes 40,000 sessions and Conv. RateConv. RateWebsite visits that ended in a purchase. Not the same as Google Ads Conv. Rate, which divides by clicks.3.5% or more is healthy reads 2%, so 800 orders. April takes 32,000 sessions and the rate holds at exactly 2% — 640 orders. At a €50 average order that's €40,000 against €32,000.
Nothing on the site changed. €8,000 went missing upstream, and only this card shows it. That's why it's the first number to check when revenue moves.
It answers the question
How many people came through the door? Orders, revenue and every rate on the Summary are a slice of this number, so one move here explains several cards at once.
Why it matters
It's the denominator. Conv. Rate on this page is purchases divided by this figure, so a traffic swing changes what that rate means without one thing on the site changing.
It also settles an argument in a single glance. Revenue down with sessions down puts the problem upstream in marketing; revenue down while sessions held puts it on the site, the offer or the checkout. Those are two different weeks of work, and nothing else on the Summary separates them.
What good looks like
There's no target count — a healthy week for one shop is a quiet one for another. Judge it against your own last 30 days and the same period last year, with your season in mind.
The shape matters more than the level. Sessions rising while Conv. Rate holds is real growth. Sessions rising while the rate falls is traffic that didn't want you, and it costs the same to buy either way.
How to improve it
| Lever | What you do | Expect | How long | Watch out for |
|---|---|---|---|---|
| Fast Add budget where the ads already work | Put more spend behind the ad sets hitting target instead of opening new ones | Sessions up roughly in line with spend | 1–2 weeks | The cheapest slice of the audience runs out first, so cost per visit climbs and Ad Spend grows faster than the sessions do. |
| Fast Mail the list you already own | Run a campaign to the segments you haven't sent to this month | A few days of extra visits | Same week | Every extra send costs unsubscribes, so the list you can mail next quarter is smaller. These are visits you already had, not new reach. |
| Slow Win the searches you nearly rank for | Rewrite the category and product pages sitting on page two for terms with real volume | Sessions that keep arriving with no cost per click | 1–2 quarters | Months of nothing first, and the work never appears in Ad Spend — so the traffic looks free when it wasn't. |
| Slow Give past customers a reason to return | Turn on restock alerts and post-purchase follow-ups so buyers come back on their own | More visits per person without new reach | 1 quarter | Sessions rise while Total Users doesn't — you're asking the same audience more often, so read it beside your subscriber numbers. |
Every lever costs something somewhere. The last column is the one to read twice.
Read it with
Traffic on its own says nothing about whether it was worth having. Orders beside it turns a count into a verdict.
Growth that converted
More visits and more of them bought. Check orders grew at least as fast as sessions — if they didn't, the new traffic converts worse than what you had.
Traffic that didn't want you
You added visits and lost orders, so the extra arrivals carry less intent. Fine for a launch week you're deliberately buying reach for, expensive as a standing setting.
Smaller and sharper
You lost the visits that were never going to buy. Genuinely good — but a rate can't keep improving on a shrinking base forever.
Failing at both ends
Fewer people arriving and fewer buying. Rule out a measurement break first, because that produces this exact picture without a customer changing their mind.
Divide spend by sessions and you have what a visit costs you. Spend up with sessions flat means you're paying more for the same door traffic, and it shows up here before it shows up in revenue — the visits change on the day, the orders they didn't produce take longer to notice.
Divide one by the other and you get visits per person. 45,000 sessions from 30,000 people is a base of first-timers; the same 45,000 from 12,000 people is a base of regulars. Those two shops should spend their budgets on completely different things.
Common misreads
Two different sites, two different audiences, two different counting rules. That's exactly why this card carries its source in the name. Adding SessionsSessionsTraffic to your listings; only lifts sales if conversion holds. to it produces a number that describes nothing.
One person can start several. A retargeting push lifts sessions while the audience stays exactly the same size — check Total Users before you call it reach.
It's dropped when All Stores is selected, because it belongs to one website. Pick a single store and it comes back.
Also called
Visits · website visits · site sessions · GA sessions
See yoursYour sessions for the period, next to the orders, revenue and ad spend that came with them.
Open Summary →