Session duration

Session duration is the amount of time a user spends in your product during a single visit. A session typically starts when a user opens the app or navigates to the site and ends when they leave or become inactive for a defined timeout period (commonly 30 minutes of inactivity in web analytics).

Session duration is an engagement depth metric that complements frequency metrics like DAU and MAU. DAU tells you how many users visit. Session duration tells you how much time each visitor spends. Both are important: a product with high DAU but very short sessions may have habitual check-in behaviour without deep engagement. A product with long sessions but low DAU may be valuable but not habitual.

The metric is context-dependent. For content platforms and educational products, longer sessions generally indicate higher engagement. Users are reading, watching, or learning. For productivity tools, the relationship is more nuanced: extremely long sessions might indicate that the tool is difficult to use or that tasks take too long. For transactional products like banking apps, short sessions are desirable because they mean users accomplish their goals quickly.

Session duration in web analytics has measurement limitations. Google Analytics measures duration based on the time between page views, which means the time spent on the last page of a session is not captured. If a user views one page for five minutes and then leaves, the session duration is recorded as zero. This is why single-page sessions (bounces) have zero duration in traditional analytics, even if the user spent significant time on the page.

Session duration is measured by tracking the time between the first and last interaction in a session. In web analytics, this is typically the time between the first and last page view or event. In native apps, this is the time between app open and app close or backgrounding.

For more accurate measurement, fire engagement events at regular intervals (for example, every 30 seconds while the user is actively interacting). This captures time on the last page and reduces the zero-duration bounce problem. GA4 does this automatically with its engagement time metric.

Session duration decomposes into the factors that keep users engaged during a visit. Understanding these factors helps product teams design experiences that deliver value within appropriate time frames.

The tree shows that session duration is driven by content depth, task complexity, and navigation efficiency. For content products, more content keeps users engaged longer. For productivity tools, the goal is often to keep sessions efficient (shorter but more productive) rather than longer. The tree helps teams understand whether long sessions reflect deep engagement or frustrated users struggling to complete simple tasks.

1. 1

   For content products: increase content depth and discoverability

   Related content suggestions, personalised recommendations, and infinite scroll keep users engaged in content products. The goal is to surface the next valuable piece of content before the user runs out of interest.

2. 2

   For productivity tools: reduce time-to-task-completion

   Productivity tools should optimise for tasks completed per session, not session length. Keyboard shortcuts, templates, batch operations, and smart defaults help users accomplish more in less time.

3. 3

   Improve page and feature load speed

   Slow loading times interrupt engagement flow. Users who wait for pages to load are more likely to abandon the session. Invest in performance to keep engagement smooth and uninterrupted.

4. 4

   Design clear navigation and information architecture

   Users who can find what they need quickly spend their session time on valuable activities rather than searching. Good navigation reduces frustration and increases productive session time.

5. 5

   Build engagement loops within sessions

   Each completed action should naturally lead to the next. Completion of one task should surface a related task. Reading one article should suggest the next. These loops keep users engaged rather than reaching a dead end.