What Are Sessions vs Users in Google Analytics?
A session is one visit. A user is the device that made it. And a user is not a person, which is the single most useful thing to understand about this tool, because several other confusions resolve the moment it lands.
A User Is Not A Person
The tool counts devices and browsers rather than human beings. Somebody who finds you on a phone at lunchtime and looks again on a laptop that evening is two users. A family sharing one computer might be one. Nothing in the reporting knows the difference.
Why it works that way. It has no other option.
Recognition depends on something being retained by the browser. There is no identity behind it, so the tool records the only thing it can observe.
What that means for your figures. Overcounting.
A hundred users is fewer than a hundred people, because some of them are the same person twice. For a considered purchase, where people look several times across devices, the effect is substantial.
Where it undercounts instead. Shared devices.
Less common and it happens. A workshop computer used by several staff appears as one visitor throughout. So does a household tablet.
Why this resolves other confusions. It explains the pattern.
Returning visitor figures looking low, conversion rates looking worse than reality and totals that never quite tally all trace back to this one fact.
How to hold it. As a proxy.
Users is a reasonable indicator of reach and a poor headcount. That distinction is enough for every decision a small business needs to make.
What A Session Actually Is
A session is one visit. The interesting part is what ends it. Visits do not finish when somebody closes a tab. They finish after a period of inactivity, which produces some counting behaviour worth knowing about.
What ends a visit. Inactivity.
Google sets the default at thirty minutes without interaction. It can be adjusted in the property settings. After that, anything further counts as a new visit.
What that produces. Split visits.
Somebody reading your page, taking a call, then coming back to it forty minutes later has generated two sessions. One person, one continuous interest, two visits in the reporting.
The version that inflates counts. Interrupted browsing.
People shop and research in fragments across a day. On a site where that is normal, session counts run noticeably ahead of anything resembling visits in the everyday sense.
What does not end a visit. Leaving the page.
Somebody going away and returning within the window continues the same session, even if they visited three other sites in between.
Why the threshold is adjustable. Different sites, different behaviour.
A site where people read for an hour and one where they check an address in ten seconds have genuinely different definitions of a visit. Most businesses should leave it alone and it is worth knowing it exists.
Which Number To Report
Both are legitimate and they answer different questions. The mistake is not choosing wrongly. It is switching between them. Worse still, mixing them in one comparison so that a rise in one looks like a rise in the other.
Users answers reach. How many found us.
Better for questions about audience size, growth over a year and whether marketing is putting you in front of more devices than before.
Sessions answers activity. How much happened.
Better for questions about volume, load and how much attention the site received. A site with fewer users and more sessions each is being used repeatedly.
The ratio is itself informative. Visits per device.
If sessions run well ahead of users, people are coming back. For a considered service that is a good sign. It is also a more meaningful measure than the returning visitor figure block four covers.
What to avoid. Comparing across the two.
Sessions this month against users last month is meaningless. It happens whenever two people build reports from the same property without agreeing definitions.
What we use with clients. One of each, stated.
A users figure for reach and a results figure for outcomes, labelled clearly. Sessions rarely earns a place in a monthly summary for a small business.
New Against Returning
The tool separates visitors it has seen before from those it has not. That sounds useful and it depends entirely on the browser retaining something, which is a considerably weaker foundation than most businesses assume.
How the decision is made. Recognition.
If the browser still carries the marker from a previous visit, the visitor is returning. If not, new. There is no other test.
What breaks recognition. Ordinary behaviour.
Clearing browsing data, using private browsing, switching from phone to laptop, using a different browser or a device that discards these markers automatically. All produce a new visitor.
Which direction the error runs. One way only.
Every one of those pushes somebody from returning into new. Nothing pushes anybody the other way, so returning is systematically undercounted rather than randomly wrong.
How large the effect is. Unknowable.
It depends on your audience, their devices and their habits. We publish no figure for it. Anybody who quotes one is guessing.
What it is still good for. Direction.
If the proportion of returning visitors is rising over a year, something real is happening, since the undercounting is roughly consistent month to month.
What it is not good for. Absolute claims.
Telling anybody that a specific share of your customers are repeat visitors is stating a figure the tool cannot support.
Which Makes Loyalty Hard To Measure Here
Put blocks one and four together and you get an uncomfortable conclusion. This tool cannot tell you reliably whether people come back, because it cannot recognise a person and it loses the recognition it does have.
What it can tell you. Device level repetition.
That a given browser returned. Useful as an indicator and not the same as knowing your customers came back.
What it cannot. Whether they bought again.
Repeat business, retention and customer lifetime value are all outside what a website measurement tool can see. It watches browsers rather than relationships.
Where the real answer lives. Your own records.
Your invoices, your booking system, your customer list. A business wanting to know its repeat rate should count customers rather than visits. That data is already in the office.
Why this matters for reporting. It gets claimed.
Retention figures from analytics appear in reports as though they described customers. They describe browsers. The gap between those two things is wide.
What to do about it. Use the right instrument.
Analytics for reach and behaviour, your own systems for loyalty and value. Neither substitutes for the other and asking one to do the other's job produces confident nonsense.
Consent Changes The Counts
What a visitor agrees to affects what can be recorded, which in turn affects how that visitor is counted. This is a factual matter about your figures rather than a legal one. We do not advise on obligations.
What happens when somebody declines. Less is recorded.
Depending on the arrangement, a visit may be recorded in a limited form or not at all. Either way the counting is affected.
The effect on returning visitors. More new ones.
Where nothing is retained on the device, every visit from that person appears as a first visit. This compounds the undercounting from block four.
Why your figures may have moved. A banner change.
Businesses that changed their consent arrangement frequently saw visitor counts shift noticeably. That is measurement moving rather than demand changing.
What we will not tell you. What to do about it.
Nothing here suggests a configuration to record more than a visitor agreed to. Nothing describes what your consent notice should cover either. That is a question for somebody qualified to answer it.
What to do practically. Note the date.
If your consent arrangement changes, write down when. Otherwise a step in the figures six months later becomes an unsolvable mystery. our troubleshooting guide covers the wider incompleteness.
Do Not Compare Across Versions
Sessions and users are calculated differently in the current version from the way the previous product did it. So any figure spanning the changeover compares two different sums. The difference between them describes the change of definition rather than your business.
What changed. The rules.
How a visit is bounded, what restarts one and how devices are counted were all revised. The words stayed the same, which is precisely why the change goes unnoticed.
What that produces. A step in the data.
Figures that move at the migration point without anything happening on the website. Businesses reported that as growth or decline. In some cases somebody claimed credit for it.
What to do about the old figures. Keep them separate.
They are not worthless, since they describe what was measured at the time. They simply cannot sit on the same chart as current ones.
Where to start your history. The changeover.
Treat it as the beginning of your comparable record. Awkward for a year and clean thereafter.
The same pattern elsewhere. Other measures.
Engagement changed at the same point, which our bounce rate guide covers. Everything sits on the Google Analytics guide.
A hundred users
is not a hundred
people.
The tool counts devices and browsers. Somebody who finds you on a phone at lunchtime and looks again on a laptop that evening is two of them. That single fact explains the low returning visitor figures, the conversion rates that look worse than reality and most of the totals that never quite tally.
How we report visitor numbers:
If you want to know how many customers came back, count invoices rather than browsers.
Every guide.
One tool.
Getting it installed and configured properly, reading the reports without being misled, measuring outcomes rather than activity, the advanced reporting and what to do when the numbers look wrong.
