User Behaviour Analysis

User Behaviour Analysis combines scroll reach, time spent, clicks and repeat visits with a screenshot of a tracked page to identify UX patterns and recommend practical improvements.

Overview

Use Analyse with AI on the User Behaviour Dashboard when the individual charts show a pattern but you need help understanding the page as a whole.

The analysis:

• reviews behaviour in 5% page-depth zones;
• compares reach, dwell, clicks and revisits;
• uses click hotspots as location clues;
• considers the visible page structure in the selected device screenshot;
• produces an executive summary and detailed zone review;
• recommends measurable page improvements;
• can send selected recommendations to the Action Board; and
• can be discussed with Ask Wheremore.

It does not change the page or publish content automatically.

Access and scope

User Behaviour Analysis is workspace-, website- and page-aware.

• The active workspace controls which websites are available.
• The selected website controls which tracked pages are available.
• A specific tracked page is required.
• The page must belong to a website accessible in the active workspace.
• Markets and search groups are not used.

Wheremore validates the workspace, website and page again when the analysis starts. A page ID from another workspace cannot be used to analyse an inaccessible page.

Before you begin

You need:

1. The website assigned to the active workspace.
2. The Wheremore pixel installed.
3. Tracking enabled.
4. The page added in Manage Session Tracking.
5. A screenshot that can be captured for the selected device.
6. Processed behaviour data for the selected page, device, audience and date range.
7. Sufficient credits for the AI analysis.

Recent visits may not appear immediately because interaction data must first be processed into the behaviour aggregates used by the report.

Quick start

1. Open the User Behaviour Dashboard.
2. Confirm the active workspace.
3. Select a website.
4. Select one tracked page.
5. Leave Referrer set to All Referrers.
6. Select All traffic, Prospects or Identified visitors.
7. Select the date range.
8. Select Desktop, Mobile or Tablet.
9. Select Analyse with AI.
10. Keep the page open while Wheremore captures the page and generates the report.
11. Review the recommendations, then expand Detailed Zone Analysis for the evidence.

Apply the filters

The analysis uses the filters that are active when Analyse with AI is selected.

Website and page

Select one website and one tracked page.

The page selector comes from Manage Session Tracking. It does not contain every URL discoverable on the live website.

Referrer

User Behaviour Analysis currently supports All Referrers only.

The aggregated scroll, click, time and revisit tables do not contain a shared session-level referrer identity. Wheremore therefore does not estimate or falsely attribute those behaviours to a selected referrer.

If a specific referrer is selected, the analysis explains that it is unavailable. Return the filter to All Referrers before trying again.

Use Referrer Pages vs Sessions or Referrer Traffic by Page when the question is specifically about traffic sources.

Audience

Choose:

• All traffic to include all qualifying behaviour;
• Prospects to include activity not classified as authenticated or identified; or
• Identified visitors to include activity classified as authenticated or identified by the tracking configuration.

Audience classification depends on the Cookie or Cookieless configuration, authenticated-user rules and exclusions set in Manage Session Tracking.

“Identified” is an audience classification. The analysis does not reveal a visitor’s private identity.

Date range

The selected range can contain between 1 and 365 days.

The analysis uses complete UTC days and ends on yesterday. It does not include today’s partial data. For example, a 30-day analysis uses the 30 complete days ending yesterday.

A longer range can provide a stronger sample, but it can also combine behaviour from different versions of the page. Use a shorter range after a major layout or content change.

Device

Choose one specific device:

• Desktop
• Mobile
• Tablet

All devices cannot be used for the AI analysis because one screenshot cannot safely represent several responsive layouts.

The selected behaviour device must match the screenshot device. Desktop behaviour is analysed against a desktop screenshot, mobile behaviour against a mobile screenshot, and tablet behaviour against a tablet screenshot.

What happens during analysis

Wheremore:

1. confirms that the workspace and tracked page are accessible;
2. captures the selected page for the selected device;
3. adds 5% page-depth bands to the analysis image;
4. retrieves scroll, click, time-spent and revisit aggregates for the selected filters;
5. groups click locations into hotspot clues where data is available;
6. asks the AI to interpret the evidence without inventing unsupported element interactions;
7. validates and formats the response; and
8. stores the completed analysis so it can be discussed with Ask Wheremore.

The process runs asynchronously. The page shows Capturing your page… and Generating your insights… while the job is active.

If a processing attempt becomes stale, Wheremore can retry it automatically. A report normally stops waiting after five minutes and shows an error if it has not completed.

Read the report

The finished report is titled Scroll Depth Analysis.

Page Optimisation Recommendations

The first section contains the strongest recommended changes.

Each recommendation can include:

Recommendations are intended to be testable. They can target one of these outcome signals:

• scroll-band reach;
• click rate;
• clicks per visitor;
• average time spent; or
• revisit rate.

Review the evidence and the live page before implementing a recommendation.

Detailed Zone Analysis

Select Detailed Zone Analysis to expand the supporting report.

Executive Summary

The Executive Summary highlights the most important patterns across the page. It should reference specific evidence rather than making a general judgement about the page.

Zone cards

Each zone card covers a 5% page-depth band with qualifying data.

What We Saw can include:

• clicks and qualifying visitors;
• click rate or clicks per visitor;
• time, dwell or revisit behaviour;
• visible content or interface elements; and
• click-hotspot location clues.

What It Means explains the likely UX interpretation while keeping observed evidence separate from inference.

Use the zone cards to understand why a recommendation was made and whether it is supported by enough data.

Interpret click evidence carefully

The report uses two related but different forms of click evidence:

• Band-level evidence shows how many clicks occurred within a 5% depth zone.
• Hotspot evidence shows where click clusters occurred within that zone using relative page coordinates.

These signals do not automatically prove that visitors clicked a particular button, completed a form, selected a plan, played a video or converted.

An exact element should only be named when the screenshot clearly isolates one target and there are no competing nearby elements. If several controls are visible, treat the hotspot as a location clue rather than exact element attribution.

Clicks at the page boundary can also come from fixed widgets, persistent controls or coordinate clamping.

Add a recommendation to the Action Board

Select Add to Board on a recommendation you want to manage as work.

1. Select Add to Board.
2. Choose a project when more than one eligible project uses the website.
3. Confirm the success message.

The Action Board item retains:

• the page and scroll zone;
• the selected device, audience and date range;
• the current state and recommended change;
• the priority;
• the supporting evidence and baseline;
• the source analysis and recommendation identity; and
• the outcome metric intended to measure the change.

Only projects connected to the analysed page’s website and accessible to the current user can be selected.

Selecting the same recommendation again does not create a duplicate item. Wheremore returns the existing item instead.

The Add to Board action does not edit the website. It creates a task for review and implementation.

Chat about the analysis

After the report completes, select Chat about this analysis to open Ask Wheremore with the analysis in context.

Useful follow-up questions include:

• Which recommendation has the strongest evidence?
• Why was this zone marked as immediate?
• Is this pattern supported by enough visitors?
• What does the click hotspot prove, and what does it not prove?
• Which change should I test first?
• How should I measure whether the change worked?

Ask Wheremore can explain the completed report, but it should not convert aggregate click clues into exact element interactions that the data does not contain.

Export the report

Select the document export control in the report header to download the displayed analysis as a text file.

The export contains the rendered report text. It does not contain the raw behaviour tables, screenshot image or individual visitor data.

Recommended workflow

1. Start with Desktop, All traffic, 30 days and All Referrers.
2. Run the analysis and read the Executive Summary.
3. Check the visitor and click evidence behind each priority recommendation.
4. Expand the zone cards for context.
5. Repeat for Mobile and Tablet.
6. Compare Prospects with Identified visitors when both groups have enough data.
7. Use Page Engagement to inspect the overlays directly.
8. Use Session Replay to understand individual journeys behind an aggregate pattern.
9. Send only well-supported changes to the Action Board.
10. Measure the selected outcome after the page change has been implemented.

Limitations

• The analysis is based on aggregated behaviour, not one visitor’s journey.
• It uses one device screenshot and cannot safely analyse combined-device data.
• Referrer-specific analysis is not currently supported.
• The screenshot is a captured representation and can differ from an older page version included in the date range.
• Responsive, personalised, third-party or dynamic content can affect visual interpretation.
• Click hotspots provide location clues, not guaranteed element identity.
• Long time spent can indicate attention, confusion, delay or an inactive tab.
• A revisit can indicate interest, comparison or difficulty; it is not automatically positive or negative.
• Small samples can produce unstable rates.
• The report does not measure conversion or revenue unless a separate source explicitly supplies that evidence.

Troubleshooting

Please select a page first

Choose a tracked page from the Pages selector. Add the URL in Manage Session Tracking if it is missing.

Please select a workspace first

Choose the correct workspace, then reselect the website and page.

Select Desktop, Mobile or Tablet

The AI analysis cannot use All devices. Choose one device so the behaviour and screenshot represent the same layout.

User Behaviour Analysis currently supports All Referrers only

Return Referrer to All Referrers. Use the referrer reports for traffic-source analysis.

Screenshot capture failed

1. Confirm that the tracked URL is valid and accessible.
2. Try the selected device again.
3. Refresh the page capture in Manage Session Tracking.
4. Check whether the page blocks automated capture.
5. Retry after images and dynamic content can load.

No scroll behaviour data found

1. Increase the date range.
2. Select All traffic.
3. Try another device.
4. Confirm the page was tracked during the selected period.
5. Check consent and exclusion settings.
6. Allow background aggregation to complete.

Analysis timed out

The browser waits for up to five minutes.

1. Keep the page open and try again.
2. Use a shorter date range.
3. Retry during lower server load.
4. Contact support if focused attempts repeatedly fail.

Analysis failed after multiple attempts

The processing job could not complete after its automatic retries. Retry once with the same filters. If the problem continues, record the workspace, page, device, audience and date range before contacting support.

A recommendation appears too specific

Check whether the screenshot isolates the named element and whether the zone card contains exact element evidence. If it only contains band-level clicks or hotspot coordinates, treat the statement as a hypothesis and verify it in Page Engagement or Session Replay.

Add to Board is unavailable

Confirm that:

• the analysis completed successfully;
• the recommendation still belongs to the active analysis;
• an accessible project is connected to the page’s website; and
• the recommendation has not expired or been altered.

Rerun the analysis if its source identity is no longer valid.

The wrong project is offered

Only projects connected to the analysed website should appear. Check the active workspace and project-to-website assignments before continuing.

Chat about this analysis does not appear

The chat control appears after the completed analysis has been stored successfully. Run the analysis again if necessary.

Frequently asked questions

Is User Behaviour Analysis the same as Page Engagement?

No. Page Engagement displays the underlying scroll, time and click overlays. User Behaviour Analysis combines those signals with a screenshot and produces an AI interpretation and recommendations.

Is it the same as Session Replay?

No. Session Replay shows an individual recorded journey. User Behaviour Analysis evaluates aggregated patterns across qualifying visitors or sessions.

Why must I select one device?

Desktop, mobile and tablet can have different layouts. The AI needs one matching screenshot so it can relate page-depth evidence to the correct visual structure.

Why can I not select one referrer?

The behaviour aggregates do not share a session-level referrer identity. Applying a specific referrer would imply precision the data does not support.

Does the analysis include today?

No. It uses complete UTC days ending yesterday.

Does a click hotspot identify the clicked element?

Not necessarily. It identifies a relative area of the page. Exact element attribution requires clearer element-level evidence.

Does a recommendation change the website?

No. The report is read-only. Add to Board creates a task but does not implement the change.

Can I discuss the report with Ask Wheremore?

Yes. Select Chat about this analysis after the report completes.

Can Wheremore measure the result of a recommended change?

Action Board items retain the selected outcome metric and baseline when measurement is available. Outcome monitoring begins only after the recommendation has been implemented and marked appropriately in the workflow.