Sessions
What it means
A visit to your own shop, as recorded by Google Analytics. A session starts when someone lands on the site and closes after 30 minutes of inactivity. It counts visits, never people — that's Total UsersTotal UsersCompare with sessions to gauge how often people return..
Show the math
Formula and a worked example
A visit runs from the first thing someone does on your site until 30 minutes pass with nothing.
Worked example. March takes 40,000 sessions. 800 end in an order, so 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 is 2%, and at €50 an order that's €40,000. April takes 32,000 sessions and the rate holds at exactly 2% — 640 orders, €32,000. Nothing on the site changed. You lost €8,000 to traffic alone.
That's why it's the first number to check when revenue moves. It sets the size of everything measured as a share of it.
It answers the question
How much traffic did the shop actually get? Orders, revenue and every rate on the page are a slice of this number, so one move here explains several of them at once.
Why it matters
Sessions is footfall through the door. You can improve everything inside the shop and still take less money if fewer people walk in — and a traffic spike works the other way, hiding a site that converts badly.
It also settles arguments fast. Revenue down with sessions down puts the problem upstream in marketing; revenue down while sessions held puts it on the site or the offer.
What good looks like
There's no universal target — a session count that's healthy for one shop is a slow week for another. Judge it against your own last 30 days and the same period last year, with your season in mind. Then read the shape: sessions rising while 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 holds is real growth, sessions rising while it falls is traffic you paid too much for.
How to improve it
| Lever | What you do | Expect | How long | Watch out for |
|---|---|---|---|---|
| Fast Mail the list you already have | Run a campaign to the segments you haven't sent to this month | Sessions up for a few days | Same week | Every extra send drives unsubscribes, so the list you can mail next quarter is smaller. These are visits you already owned, not new reach. |
| Fast Raise budget where the ads already work | Add spend to the ad sets hitting target instead of opening new ones | Sessions up roughly in line with spend | 1–2 weeks | Cost per click climbs once you push past the cheapest slice of the audience, so Contribution Margin thins while traffic grows. |
| Slow Earn the searches you nearly rank for | Rewrite the category and product pages sitting on page two for terms with real volume | Sessions up month after month with no cost per click | 1–2 quarters | Months of nothing, and the content spend never appears in Ad Spend — the traffic looks free when it wasn't. |
Every lever costs something somewhere. The last column is the one to read twice.
Read it with
Sessions counts visits. On its own it can't say whether more visits means more people, or the same people coming back more often.
Growth that held its quality
More visits and a better rate at the same time. That's usually one source doing something right rather than the whole site improving, so go and find which.
Traffic you paid too much for
The extra visits carry less intent than the ones you had. 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. The rate improved by shrinking, which is real — but a rate can't grow forever on a falling base.
Failing at both ends
Fewer people arriving and fewer of them buying. Rule out a tracking break first, because a broken tag produces exactly this picture.
Divide one by the other and you get visits per person. 45,000 sessions from 30,000 people is 1.5 visits each, mostly first-timers. The same 45,000 from 12,000 people is a base of repeat visitors, and the two call for opposite budgets.
One counts arrivals, the other counts arrivals that landed. When sessions climb and engaged sessions don't move, you've bought volume rather than attention — a gap that shows up here weeks before it shows up in revenue.
Common misreads
One person can start ten sessions in a week, and a retargeting push lifts sessions while the audience stays exactly the same size. Check Total Users before you call it reach.
They never do. The two count with different timeout rules and different bot filtering. Use each one against its own history.
A step change in a day is nearly always measurement. A tag removed from a template, a new consent banner or a redirect that drops parameters will wipe the count while the real visits carry on arriving.
Also called
Visits · site sessions · GA sessions
See yoursYour sessions for the period, the change against the period before, and the sources that made them up.
Open Google Analytics →Sessions
What it means
A visit to one of your Amazon listings, as counted by Amazon's Sales & Traffic report. These people are browsing Amazon, not your shop — you never see them arrive and you can't retarget them. Amazon counts a separate metric, Page ViewsPage ViewsHigher means more product visibility., for the pages viewed within a visit.
Show the math
Formula and a worked example
A session is one visitor's activity on your listings within a 24-hour window, as Amazon defines it.
Worked example. A SKU takes 6,000 sessions in March and ships 720 units, so Unit Session %Unit Session %How well your listings turn visits into sales. Counts units, so a 3-unit order counts as 3.12% or more is healthy is 12%. In April sessions fall to 4,000 while the rate holds — 480 units. The listing didn't get worse; fewer people saw it.
Sessions on Amazon are mostly won by search rank and advertising, not by anything on your own site.
It answers the question
How many shoppers actually reached your listing? Unit sales on Amazon are this number multiplied by how well the listing converts — and only one of those two is something you control day to day.
Why it matters
On Amazon you're renting attention. Sessions tell you how much of it you got, and they move for reasons outside your shop entirely: a competitor outbidding you, a search-rank slip, or going out of stock.
It's also the denominator underneath Unit Session %Unit Session %How well your listings turn visits into sales. Counts units, so a 3-unit order counts as 3.12% or more is healthy, so a change here quietly moves that rate even when the listing is untouched.
What good looks like
No universal target — it scales with your category and how many SKUs you list. Track each SKU against its own trend. A sustained fall with no price or stock change usually means search rank, and that's the thing to go and check.
How to improve it
| Lever | What you do | Expect | How long | Watch out for |
|---|---|---|---|---|
| Fast Fix anything that suppresses the listing | Clear stock-outs, pricing errors and policy flags that push a listing down or hide it | Sessions recover within days | Same week | It restores what you had rather than adding to it — a fix, not growth. |
| Fast Bid on the terms you already convert on | Raise bids on search terms with sales, using the search-term report | Sessions up while the budget lasts | 1–2 weeks | Bought sessions stop the day the budget does, and ACOS rises with the bid. |
| Slow Earn organic rank | Improve title, images and reviews so the listing ranks for its main terms without paying | Sessions climb and stay | 1–2 quarters | Slow, and hard to attribute — rank moves for reasons you can't see. |
Every lever costs something somewhere. The last column is the one to read twice.
Read it with
Sessions tell you how many people arrived. Whether the listing deserved them is a different number.
Sessions falling with the rate steady is a traffic problem — go and look at rank and ads. Sessions steady with the rate falling is a listing problem — price, images, reviews or a competitor undercutting you. The two need completely different weeks of work.
Losing the Buy Box doesn't reduce sessions — people still arrive — but the order goes to another seller. Sessions holding while units fall, with Buy Box % slipping, is that exact story.
Common misreads
They can't. Different audiences on different sites, counted with different rules. Adding them produces a number that describes nothing.
On Amazon, sessions usually track visibility rather than demand. Check search rank, stock and whether a competitor started advertising against you.
Also called
Amazon sessions · listing visits
See yoursSessions by marketplace and by SKU, next to units ordered and the rate between them.
Open Amazon Summary →