TL;DR, Quick Answer
7 min readA session is one continuous group of a visitor's interactions on a site, opened by the first event and closed by inactivity, a new day boundary, or a change of campaign source. The closing rules differ between tools, so the same traffic can produce two sessions in Google Analytics 4 and four in Matomo. Sessions are sittings, not people, and most rate metrics divide by them.
What is a session in web analytics?
In web analytics, a session is one continuous group of a visitor's interactions on a site, opened by the first event and closed by inactivity, a new day boundary, or in some tools a change of campaign source. Analytics tools use the session as the container that holds pageviews, clicks, scrolls and conversions for a single sitting. Every product writes its own closing rules, so the number you read depends on which product produced it.
Does a session mean the same thing as a login session?
No. In authentication, a session is the server-side or cookie-backed state that keeps a person signed in until they log out or the token expires, which is a separate mechanism from the analytics unit described here. Everything below uses the analytics meaning.
When does a session start and end?
A session starts on the first tracked event from a visitor who has no session currently open, and ends when one of the tool's closing rules fires. Three closing rules are in common use: an inactivity timer that resets on every event, a calendar day boundary in the site's configured timezone, and a change in the campaign parameters attached to the URL. Before comparing your numbers with anyone else's, find those three rules in your own tool's reference and write down which ones it applies, starting with the inactivity timer.
How does GA4 close a session?
Google Analytics 4 closes a session after 30 minutes of visitor inactivity by default. Google's About Analytics sessions reference states that "a session ends or times out after 30 minutes of user inactivity", and the timer can be raised in the interface to a maximum of 7 hours and 55 minutes. That page lists inactivity as the closing condition and describes no midnight cutoff, which is one of the breaks from Universal Analytics, where a session expired at midnight in the view's timezone.
Campaign handling changed in the same migration. Google's Universal Analytics to GA4 comparison puts it directly: "In Universal Analytics, a new campaign will start a new session regardless of activity. In Google Analytics 4, a new campaign does not begin a new session." A visitor who arrives from an ad, then clicks a newsletter link 10 minutes later, produced two sessions in Universal Analytics and produces one in GA4. All three Google rules above were checked in September 2026.

How do other analytics tools close a session?
Every major tool starts from a 30-minute inactivity timer and then adds its own extra conditions on top.
| Tool | Inactivity timeout | New calendar day | New campaign source | Other limits |
|---|---|---|---|---|
| Google Analytics 4 | 30 minutes, adjustable to 7h55m | not documented as a closing rule | does not open a new session | none documented |
| Universal Analytics (retired) | 30 minutes | closes at midnight, view timezone | opens a new session | none documented |
| Adobe Analytics | 30 minutes | not documented as a closing rule | not documented as a closing rule | 12 hours of unbroken activity, 2,500 hits, or 100 hits in the first 100 seconds |
| Matomo | 30 minutes, set by visit_standard_length | opens a new visit after midnight, site timezone | new campaign parameters or ad click IDs open a new visit | none documented |
| Plausible | 30 minutes | not documented as a closing rule | not documented as a closing rule | none documented |
The Adobe limits come from its Visits metric reference, which says "if more than 30 minutes lapse between hits, a new visit begins" and adds the 12-hour, 2,500-hit and 100-hits-in-100-seconds caps. The Matomo rules come from its visit definition FAQ, where the 30-minute default is set by visit_standard_length in seconds. Plausible's metrics definitions state that "a visit starts when a visitor first lands on your page and ends when no action is taken for 30 minutes". Where a cell says not documented, the vendor's own reference lists no such rule, which is weaker evidence than a stated guarantee.
Why does the same traffic produce different session numbers?
Different closing rules split the same event stream at different points. Take one visitor with six events in a single evening, in the site's timezone:
| Time | Event | Gap since last event |
|---|---|---|
| 21:40 | lands from an organic search result | first event |
| 22:05 | views a second page | 25 min |
| 22:15 | clicks a newsletter link carrying utm_campaign=autumn-digest | 10 min |
| 22:35 | views a fourth page | 20 min |
| 23:50 | views a fifth page | 75 min |
| 00:10 | views a sixth page | 20 min |
A tool that closes only on inactivity breaks once, at 23:50, and reports two sessions. Matomo breaks three times, at 22:15 for the campaign change, at 23:50 for the timeout, and at 00:10 for the day boundary, and reports four visits. One visitor, six events, a 2x difference in the headline number, and no difference in the traffic. The same split explains why one person on a phone and then a laptop registers as two sessions and two visitors, which is what stitching a visitor's sessions together sets out to correct.

Is a session count a people count?
A session count is a count of sittings, not of people. One visitor who returns on Monday, Wednesday and Friday adds three sessions and one visitor, so sessions always run ahead of visitors on any site with returning traffic:
sessions per visitor = sessions / visitors
For a site reporting 12,400 sessions and 5,200 visitors in a month, that ratio is 12,400 / 5,200 = 2.38 sessions per visitor. Read the ratio as return frequency, and never report a session figure with a label like "people reached" or "users" attached to it.
Which metrics divide by sessions?
Bounce rate, pages per session, average session duration and the standard session-based conversion rate all use the session count as their denominator, which makes the boundary rules above a direct input to each one. Work the conversion rate as an example:
conversion rate = conversions / sessions
With 310 conversions and 12,400 sessions, the rate is 310 / 12,400 = 0.025, or 2.5 percent. Move the same traffic and the same 310 conversions into a tool that splits on campaign changes and midnight and reports 18,600 sessions, and the rate reads 310 / 18,600 = 0.0167, or 1.67 percent. Nothing about the site changed. Pin one tool as the system of record for every rate metric, and note the denominator next to the number in any report that leaves your team.
Flowsery
Start FREE Trial
Real-time dashboard
Goal tracking
Cookie-free tracking
Flowsery records every user session and its built-in analytics reports real-time traffic, funnels, goals and user journeys with no data sampling, so the counts cover the full event stream instead of an extrapolation from a subset. Sessions carry the pageviews and events that make each rate metric meaningful, so a session boundary that moves takes every downstream percentage with it.
Frequently asked questions
How long is a session before it times out?
Thirty minutes of inactivity is the default in Google Analytics 4, Adobe Analytics, Matomo and Plausible. The timer resets on every tracked event, so a visitor reading one page for two hours without scrolling or clicking still times out at the 30-minute mark from their last event. GA4 allows the timer to be raised to 7 hours and 55 minutes.
Does a session end at midnight?
That depends on the tool. Universal Analytics closed every session at midnight in the view's timezone, and Matomo opens a new visit when a visitor continues browsing past midnight in the site's timezone. Google's GA4 session reference describes inactivity as the closing condition and lists no midnight cutoff.
Does clicking a new ad start a new session?
In Google Analytics 4, no. Google's migration documentation states that a new campaign does not begin a new session in GA4, reversing the Universal Analytics behavior where a campaign change closed the open session and opened a fresh one. Matomo does open a new visit when a visitor arrives with new campaign parameters or ad click IDs.
Why does my tool report more sessions than visitors?
One visitor produces a new session on every return trip, so any site with repeat traffic reports more sessions than visitors. Dividing sessions by visitors gives return frequency: 12,400 sessions across 5,200 visitors is 2.38 sessions per visitor. A ratio close to 1.0 means almost nobody is coming back.
Can I change the session timeout?
Google Analytics 4 exposes the timeout in the interface up to a ceiling of 7 hours and 55 minutes. Matomo On-Premise sets it through visit_standard_length in the config file, in seconds. Changing the timeout changes every session-based metric on the site and breaks comparability with your own historical data, so move the value once and record the date you moved it.
Is a session the same as a session replay?
No. A session is the counted unit in your analytics reports, while a session replay is the reconstructed recording of what happened inside one of those sittings. The two share a boundary definition in tools that record and count in the same system, which is what lets a replay be opened directly from the session row that flagged a problem.
What actually gets counted inside one session?
A session is the container analytics tools use to hold everything a visitor does in one sitting: pageviews, clicks, scrolls and conversions. It opens on the first tracked event and stays open until a closing rule fires, so every action logged before that point belongs to the same session. Two visitors who each view three pages and click once produce sessions with the same contents, even when the tools disagree on where the boundary falls.
What limits does Adobe Analytics add on top of the 30-minute timeout?
Adobe closes a visit after 30 minutes of inactivity like the other major tools, then adds three more caps: 12 hours of unbroken activity, 2,500 hits, or 100 hits within the first 100 seconds. Whichever limit hits first closes the session, so a visitor who stays engaged past the inactivity window can still get cut off by the hit count. Adobe's own Visits metric reference lists these caps, and no other tool in the comparison documents anything like them.
How does Plausible define when a session starts and ends?
Plausible's metrics definitions state that a visit starts when a visitor first lands on a page and ends when no action is taken for 30 minutes, the same inactivity window every other major tool defaults to. Plausible documents no midnight cutoff and no campaign-change rule, so an evening of browsing with a mid-visit ad click stays inside one Plausible session unless the visitor goes quiet for half an hour.
What's the difference between a session and a pageview?
A pageview is one loaded page. A session is the container that holds every pageview and event a visitor generates before a closing rule fires, and metrics like pages per session divide one by the other. A session boundary that shifts changes both numbers at once, since sessions carry the pageviews that make the ratio meaningful.
Was This Article Helpful?
Let us know what you think!
See us more often in Google
One click marks Flowsery as a preferred source, so our articles sit higher in your Top Stories, AI Mode, and AI Overviews.
Before you go...
Flowsery
Revenue-first analytics for your website
Track every visitor, source, and conversion in real time. Simple, powerful, and cookie-free.
Real-time dashboard
Goal tracking
Cookie-free tracking
Related Glossary Terms


How Session Replay Works, and What It Cannot See
A session replay rebuilds a visit from DOM mutations and input events, not video. See what it captures, what masking hides, and how it differs from heatmaps.


Two Numbers Hide Behind One Drop-off Rate
Every funnel produces two drop-off rate numbers, one per step and one end to end, and teams quote them interchangeably. A worked table separates them.


The Setup Choices Behind Every Funnel Analysis
Three setup choices decide what funnel analysis reports: step sequencing, the conversion window, and whether the funnel counts users or sessions.


Five Ways to Calculate Net Revenue Retention From One Dataset
One net revenue retention formula, five defensible variants: the same cohort returns 84.0%, 104.5%, 108.3%, 109.5% or 110.3% depending on window and base.
Related Articles


Answer the Monday Question With a Grok Bot Template for Analytics Reporting
A Grok bot template for analytics reporting turns a weekly traffic question into one API call. The breakdown endpoint takes 25 dimensions, which is what lets a chat answer follow-ups a dashboard cannot.


Answer Traffic Questions With a Grok Bot Template for Web Analytics
A Grok bot template for web analytics answers traffic and session questions in chat, including the issues found in recordings. The skill text, the weekly routine, and the one call that changes data.


A Practical Guide to E-Commerce Analytics
Conversion rate, revenue per visitor, average order value and checkout abandonment: the store metrics worth tracking, and the ones that only add risk.

