Explorations are the advanced analysis area of GA4, found under Explore in the left-hand menu, where you build your own tables, funnels and journey diagrams from event data instead of relying on the fixed standard reports. They answer the questions the standard reports were not designed for.
How explorations work
Each exploration has three panels. On the left you pick your variables: the date range, the segments, and the dimensions and metrics you want to work with. In the middle you choose a technique and drag variables into it. On the right the results appear, and you can right-click a row to drill further or build a segment from it.
The techniques available at the time of writing (October 2026) are:
- Free form: a flexible table or chart, for example landing pages by device with key events alongside.
- Funnel exploration: steps you define, such as product page, basket, checkout and purchase, with the drop-off at each.
- Path exploration: the sequence of pages or events people follow from a starting point, or the route that led to an end point.
- Segment overlap: how up to three segments intersect, such as mobile users, returning users and purchasers.
- Cohort exploration: how groups who first arrived in the same week behave over the following weeks.
- User lifetime: behaviour and value across the whole time GA4 has known each user.
- User explorer: the event stream for individual anonymous users.
Explorations are private to the person who made them unless shared, and shared copies are read-only for everyone else on the property.
Why it matters
The standard reports cover the common questions: where traffic comes from and which pages are popular. The questions that change decisions are usually more specific. Which landing pages bring paid search visitors who go on to request a quote? At which checkout step do mobile shoppers drop out? Do people who read a guide come back within a month? Explorations answer those without exporting data or paying for another tool.
Common mistakes
- Leaving data retention at two months. Explorations can only look back as far as the property’s data retention setting, which is two months on a new property. Change it to 14 months under Admin, Data collection and modification, Data retention.
- Ignoring sampling. Large explorations on busy properties can be based on a sample of the data. The indicator at the top of the exploration tells you; shorten the date range if precision matters.
- Mixing scopes Such as user metrics against an event-scoped dimension, and then adding up rows that overlap.
- Using explorations for routine reporting. They are built for investigation. For a monthly report that someone else reads, customise the standard reports or build a dashboard in Looker Studio.
- Being caught out by thresholding. GA4 can withhold rows with small user counts to protect privacy, most often when demographic data is involved, which can make low-volume segments vanish. An indicator at the top of the exploration shows when this has happened.
How to act on it
- Check the data retention setting today, before you need the history.
- Start each exploration from one question you actually need answered, then pick the technique: free form for comparisons, funnel for drop-off, path for journeys.
- Build a handful of segments you will reuse, such as paid traffic, mobile users and people who triggered a key event.
- Name explorations clearly, and share the ones colleagues need.
- When a result looks odd, check sampling, thresholding and scope before acting on it.
Explorations are where I look for the reasons behind the numbers when reviewing a business’s marketing as part of digital marketing strategy and consulting.