Filters

A filter narrows your entire dashboard to a slice of visitors. Every panel, the graph, the map, the KPIs, and the session list all reshape around the same condition at once, so you can ask a precise question and read the whole answer in one view. Open the filter editor from the layers icon in the top bar, or press F anywhere on the dashboard (see keyboard shortcuts).

How a filter is built

Pick a category on the left, then a dimension under it, then one or more values. The categories group the 25 dimensions you can filter on (UTM tags expands into the five individual UTM parameters):

• URL: Page, Entry page, Exit page, Outbound link.
• Behavior: Goal and Signal.
• Acquisition: Channel, Source, Campaign, UTM tags, Referrer URL.
• Location: Continent, Country, Region, City.
• Device: Browser, Browser version, OS, OS version, Device, Screen.

The example below stacks three conditions: page contains /blog, source is Google or Bing, and country is Germany. The badge next to a dimension in the picker shows how many values that dimension currently holds.

When you are done, Apply filter puts the condition live, Clear all resets the editor, and + Save as segment stores the whole condition for later (see Saved segments below).

Operators

Most dimensions offer four operators:

• is: an exact match.
• is not: everything except the value.
• contains: the value appears anywhere in the field. Case-insensitive, useful for paths and URLs.
• does not contain: the value appears nowhere in the field.

Goal rows are the exception: they use did complete and did not complete instead, because a goal is something a visitor either did or did not do. Signal rows are membership only, so they use is.

Excluding traffic

Set a row's operator to is not or does not contain and the condition excludes instead of includes. Negated values render as red chips, so an excluding condition is always visually distinct from an including one, and you can mix both kinds in the same filter.

The most common use is cleaning your numbers without touching the tracking snippet. The example below removes admin and preview pages, an internal test source, and Direct traffic in one filter:

Matching several values (OR)

Add more than one value to a single row and they are combined with OR. Source is google, bing matches visitors from either. Country is not US, CN, IN excludes all three at once. There is no limit on the number of values inside one row, because they compile to a single list, which is cheap no matter how long it gets.

Multi-value rows stack like any others. The example below combines three of them: country is Germany, Austria, or Switzerland, device is mobile or tablet, and browser is Chrome, Safari, or Firefox.

Stacking conditions (AND)

Add a second row for the same dimension, or for any other dimension, and the rows combine with AND. The editor example at the top of this page reads as: visited a page containing /blog, and arrived from Google or Bing, and located in Germany.

This is the rule for the whole editor: values inside one row are OR, separate rows are AND. That covers any question you can phrase as "all of these conditions, where each condition allows a few values." You can stack up to five conditions per dimension.

Once applied, the filter icon in the header shows a badge with the number of active filter values. Every KPI, chart, and panel recalculates against all active conditions at once. Click the icon any time to review or edit the conditions; negated conditions show as red chips inside the editor:

Filtering by goal completion

Under Behavior, Goal lets you filter to visitors who completed a goal. When at least one goal is set to did complete, the breakdown panels switch their visitor column to Conversions with a per-row conversion rate, so you can see which sources, pages, and countries convert best. Stack several did complete goals to require all of them. A did not complete goal is a population filter: it narrows who you are looking at without turning the view into a conversion report. See Goals for how goals are defined.

Filtering by signal

Under Behavior, Signal filters to visitors who hit a friction point: rage clicks, dead clicks, copied text, script errors, form abandons, or prints. It is the fastest way to ask "show me everyone who got stuck" and then read where they came from and where they dropped off. For the detail behind each category, expand a signal in the Sessions panel. See Signals.

Cross-filtering from panels

You do not have to open the editor to start a filter. Click any row in a panel, a source, a page, a country, a goal, a signal, and the dashboard filters to it instantly. Clicking a second value on the same panel adds it to that condition (OR). The editor is where you go to add the precise, stacked AND conditions that a single click cannot express.

The condition is written into the page URL as it is applied, for example ?filters=source:google, so a cross-filtered view can be bookmarked or shared like any other:

Saved segments

Once you have a filter you return to, save it as a segment from the editor. Segments are stored server-side per site, not in your browser, so they are available wherever you sign in. Each one reloads the whole condition in one click. They are the right home for the multi-condition filters you check often, like "blog readers from Europe who did not sign up."

Saved segments are listed at the bottom of the filter editor, with the number of conditions each one carries:

Shareable links

The active filter is written into the page URL, so the view is fully shareable. Copy the address and anyone with access to the site lands on the same filtered dashboard. Old links keep working unchanged, so bookmarks never break.

Frequently asked questions

Can I export the filtered view?

Yes. Both the PDF report and the CSV export respect the active filters, so you export exactly the segment on screen, not the whole site. See Export Data and Reports.