What a session measures
A session groups everything one person does on your site into a single visit. It opens when the first page view or event is recorded and closes after a stretch of inactivity that the analytics tool defines. Come back after that gap and the tool opens a fresh session, even though the browser and the person have not changed.
Google Analytics 4 also closes a session when the day rolls over in the property’s reporting time zone. The inactivity timeout itself can be adjusted in the admin settings, which is worth knowing if your pages ask for long reading or carry a slow booking form where people naturally sit still for a while.
Why sessions matter
Almost every headline traffic number is built on sessions rather than on people. Conversion rate is usually conversions divided by sessions, channel reports count sessions, and a year-on-year traffic comparison compares them. Once you hold on to the fact that a session is a visit and not a visitor, those numbers stop being confusing. Sessions rising while users stay flat means people are returning more often, which is a very different story from reaching a new audience.
GA4 adds engaged sessions, which only count visits that lasted, went beyond one page, or triggered a conversion. That distinction is more useful than raw sessions when you are judging whether a page is doing its job, because it separates a real read from an instant exit.
Where sessions go wrong
Comparing session counts between tools is the most common trap. Analytics platforms rarely define the timeout, the reporting time zone and the rules for restarting a visit in the same way, so their totals differ even when both are installed correctly, and neither of them is wrong.
The other problem is inflation from your own side. Staff browsing the site, developers testing a change and uptime monitors all create sessions, and on a small site that noise is easily large enough to change the picture. Arriving from a link that carries no UTM parameters tends to open a new session filed as direct traffic, which quietly moves credit away from the campaign that earned the visit.
How to act on it
Set the property up so the counts mean something. Exclude internal and developer traffic, confirm the reporting time zone matches the market you actually sell to, and adjust the inactivity timeout if your pages demand long attention. Then pick one tool as the source of truth and stop trying to reconcile the others against it.
Use sessions for volume and trend, not for judging people or value. If the question is whether marketing is working, follow enquiries and sales through a properly configured analytics and tracking setup and treat sessions as the denominator sitting underneath them. A month with fewer sessions and more enquiries is a better month, and nothing in a sessions chart will tell you that on its own.
