Connect data sources

Connected data sources are persistent organisation-level connections. UX Robot currently supports Google Analytics 4 and Google Search Console. Journey-specific browser and Microsoft Clarity evidence is gathered by autopilot after the journey brief is saved.

Before starting:

• Sign in to the Google account that can read any GA4 property or Search Console site you want to include.
• Confirm the GA4 property represents the same website as the UX Robot organisation.
• Confirm Search Console lists the correct domain or URL-prefix property.
• Allow the read-only permissions requested by UX Robot. Removing a requested scope can leave that connection incomplete.
• If your organisation restricts third-party apps, ask the Google Workspace administrator to approve UX Robot before connecting.

A new account can connect both during set-up. Select Connect beside each one on the Google data step; once you choose the property or site on Data Sources, UX Robot takes you back to set-up, and a banner there offers Back to set-up at any time.

1. Open the organisation and select Data Sources.
2. Find Google Search Console and select Connect.
3. Choose the Google account with access and approve the requested permission.
4. After returning to UX Robot, choose the correct site when more than one verified property is available.
5. Confirm the selected site shown on the Data Sources page matches the organisation website.

On every plan, each page audit shows the searches that bring people to the audited page, with their clicks, impressions, click rate and average position. On Enterprise, Search Console evidence also helps explain organic demand, search intent, content gaps, devices and countries, high-impression pages, click-through performance, and changes between periods. A domain property and a URL-prefix property can contain different coverage, so select deliberately.

1. On Data Sources, find Google Analytics 4 and select Connect.
2. Choose the authorised Google account and approve read-only access.
3. Choose the GA4 property for the organisation. Property names can be similar, so check the account and property identity carefully.
4. UX Robot attempts to retrieve the property’s key events. Review the selected property shown after the connection is saved.

GA4 is required to start a journey, because every finding is ranked by it. It supplies traffic, engagement, events, conversions, device patterns, landing pages, and other reports scoped to the journey’s evidence period. You can save a journey brief before GA4 is connected; autopilot starts once it is. GA4 also suggests journeys worth testing, puts each page change on the page timeline next to before and after figures, and shows where open issues meet the most visitors on the Impact page.

If goals should align to a GA4 key event, return to the organisation’s Goals section after connecting and update the goal’s analytics event.

Evidence is global by default. UX Robot does not infer a geographic audience from an organisation name, profile country, domain, currency, timezone, project name or website content. Add an exclusion only when traffic from that location should not be part of the evidence population, such as clearly identified automated traffic from a data centre.

1. On Data Sources, find Evidence scope and select Country exclusions.
2. Search for and select each country to exclude.
3. Select Save country filter. The excluded countries are then listed in Evidence scope.

The saved scope applies consistently to new geography-aware evidence, including GA4, Search Console and Microsoft Clarity. GA4 and Search Console apply the exclusions while aggregating their reports. Autopilot passes the same global-with-exclusions scope to Clarity and must omit mismatched behaviour evidence and record a limitation when Clarity cannot apply it exactly.

Exclusions apply to future evidence collections and new autopilot runs. An active run keeps the immutable scope saved when it started, while reports already stored keep the exclusions that were in force when they were generated. Expect counts to fall the first time a filter is applied. That is the excluded traffic, not a decline in real visits.

Your team visits its own website, and every journey with a step your team completes, such as signing up with a test account, records that sign-up in GA4. Your agent runs the journey on your team’s computer, so its test visits come from your connection too. Left in, they look like real visitors and real completions of your goals. GA4’s internal traffic filter removes them by IP address, so real customers who sign up at the same time are still counted.

You need the Editor role on the GA4 property. In GA4:

1. Find your team’s public IP address, for example by searching “what is my IP” from the connection your team uses.
2. Open Admin, then Data streams, and choose the website’s web stream.
3. Select Configure tag settings, then Show more, then Define internal traffic.
4. Select Create, name the rule (for example Team), and leave the traffic_type value as internal.
5. Under IP addresses, choose IP address equals for one fixed address, or IP address is in range (CIDR notation) for a range, enter the address and select Create. Add a condition for each office or home connection your team tests from.
6. Back in Admin, open Data filters and choose Internal Traffic. Check that it excludes traffic_type internal, change its state from Testing to Active and save.

Microsoft Clarity records your team too: on a smaller website, one person testing can make up a fifth of a page’s sessions. In your Clarity project, open Settings, then IP blocking, and add the same addresses. Clarity stops recording those visits from then on.

Select Reconnect when a token has been revoked, permissions have changed, the Google account is wrong, or the source displays an error. Complete the Google flow again and reselect the required site or property.

• No sites or properties appear: confirm the chosen Google account has direct access, then reconnect.
• The wrong item was selected: use the source’s choose-site or choose-property action and save the correct option.
• GA4 key events cannot be listed: reconnect to refresh access. You can still correct the property selection before collecting reports.
• Data API error: make sure the Google Analytics Data API is enabled for the relevant Google Cloud project and the account can read the property.

Reconnect before starting a new analysis if the source is in an error or disconnected state. Reports already stored on an older analysis remain part of that historical analysis.