Conversion Performance Analysis

Conversion Performance AI Analysis reviews a selected conversion path step by step, then combines those reports into an end-to-end journey analysis.

It uses conversion progress, form activity, page behaviour, configured conversion zones, screenshots and attribution context to identify the most useful CRO actions.

Overview

Use Analyse with AI on the Conversion dashboard when you need to move from reporting to diagnosis.

The analysis can help answer:

Where is the main loss between path entry, form start and completion?
Which conversion step should be investigated first?
Are visitors reaching a conversion zone without progressing?
Do the available page and device signals suggest a motivation, trust, clarity or friction problem?
Is an apparent issue more likely to be a measurement mismatch that needs controlled validation?
Which changes should be sent to the Action Board?

Wheremore analyses every configured step before producing the journey report. It does not edit the website or conversion path automatically.

Before you begin

You need:

Interaction tracking installed and enabled for the website.
Relevant pages included in Manage Session Tracking.
An active conversion path with ordered steps.
Conversion zones assigned to the relevant path steps.
Working selectors or form-submission tracking for the devices being measured.
Processed conversion, funnel, behaviour and attribution data.
Sufficient credits for the AI analysis.

The selected path must belong to a website available in the active workspace.

Quick start

Open the Conversion Dashboard.
Select the date range.
Select the website.
Select a conversion path.
Choose the campaign and audience scope.
Select Analyse with AI.
Keep the page open while the step reports are generated.
Review the Priority Conversion Improvements for each step.
Review Journey Priority Improvements and expand Detailed Journey Analysis.
Add supported recommendations to the Action Board or use Ask Wheremore about this CRO analysis.

If All Paths is selected, Wheremore analyses the accessible paths one after another.

Reporting scope

The analysis uses the active Conversion dashboard filters when it starts.

Date range

The selected period supplies the conversion and behavioural evidence used by the report. The path's active schedule can shorten the effective range.

Recent activity is not final until the relevant background processing has completed.

Website and conversion path

A website and a valid path are required. The path must:

belong to the selected website;
be available to the signed-in user in the active workspace; and
contain at least one ordered step.

Campaign and channel

Conversion truth and available attribution context honour the selected campaign and channel scope.

Some supporting page behaviour, zone exposure or journey evidence can come from a broader path, date, audience and device population. When this happens, the report should identify it as broader context rather than attributing it to the selected campaign or channel.

Audience

Prospects includes sessions recorded as unauthenticated.
Identified Visitors includes sessions recorded as authenticated or identified by the configured audience rules.
All Traffic includes both.

Audience is separate from Cookie or Cookieless identity.

Device

When a specific device is active, Wheremore uses the matching device evidence and screenshot where available.

Desktop, mobile and tablet zones are assessed independently. An unconfigured device does not automatically mean tracking is broken.

Markets

Markets and search groups do not apply to Conversion Performance AI Analysis.

What happens during analysis

Wheremore:

validates the workspace, website, path and ordered steps;
creates a separate asynchronous analysis job for each step;
retrieves the applicable conversion, behaviour, zone and attribution evidence;
captures the available device screenshot and conversion-zone overlay;
generates and validates the step report;
repeats the process for every configured step; and
combines the completed step jobs into a separate journey analysis.

Each step is labelled Analysing, Complete or Failed. A failed step can be retried individually. The journey analysis starts only when a complete report is available for every path step.

Use Collapse All or Expand All to manage a long report.

Read a step analysis

Each CRO Step Analysis is based on the selected step's position in the path.

Analysis context

The report header can show:

Step Entrants
Path or page entrants
Date range
Audience
Cohort
Device
Referrer
Conversion type
Value type
Target rate

Priority Conversion Improvements

These are the strongest recommended changes for the step.

Each card can include:

the current state;
the recommended change;
supporting evidence and justification;
suggested copy or UX direction;
priority; and
Add to Board.

Treat the recommendation as a testable hypothesis, not an automatic instruction to change the page.

Executive Summary

The Executive Summary describes the most important findings for the step.

Cohort Behavioral Analysis

This compares available progressor and drop-off behaviour, including early engagement signals and engagement gaps.

For an intermediate step, Wheremore uses terms such as step progressors or next-step reachers. Converter and completed session are reserved for sessions that completed the full path.

Conversion Zone Visibility

This explains where the configured zone sits, whether sessions scrolled far enough to reach it and where page drop-off occurred.

Zone reach means scroll exposure. It does not mean the visitor clicked, focused, typed into or submitted the element.

Traffic Source Performance

This provides channel and device context where supported. It should not replace the authoritative path-entry, form-start and completion totals.

Success Pattern Analysis

This summarises patterns associated with path completions, step progression and drop-off.

Small samples are directional. They should not be treated as proof that a behaviour caused conversion or abandonment.

Step Analysis and Context-Specific Analysis

These sections combine performance summaries, observed behaviour, warning signals, device evidence and temporal patterns.

The report separates observations from interpretation. Expand the detailed cards before deciding what to change.

Read the journey analysis

The CRO Journey Analysis combines the completed step reports with path-level conversion truth.

Journey Priority Improvements

These cards identify the most important end-to-end actions. They should prioritise the measured leak:

path entrants;
form starts;
completed sessions.

Where measurement validation is also needed, it should normally appear as a checklist within the main CRO action rather than as a duplicate technical recommendation.

Detailed Journey Analysis

Expand this section to review the available:

funnel bottleneck summary;
journey flow and step progression;
stage or category transitions;
device, channel and campaign context;
journey complexity;
measurement evidence; and
path-wide recommendations.

For a one-step path, the report focuses on entry, form start, completion, channels and exits. It does not invent cross-step behaviour.

Understand the source-of-truth metrics

The analysis treats these conversion metrics as authoritative:

Path entrants: qualifying sessions that entered the selected path.
Form starts: qualifying sessions that began the configured form or conversion interaction.
Step sessions: sessions that produced accepted progression for a specific step.
Completed sessions: sessions that completed the full path.
Path conversion rate: completed sessions divided by path entrants.
Form-start rate: form starts divided by path entrants.
Form-completion rate: completed sessions divided by form starts.

Attribution tables explain channels, campaigns and journey shape. Their counts can be smaller than the conversion-truth population and must not replace these totals.

Interpret zone and form evidence safely

Zone reach is not interaction

Zone exposure only shows that a session scrolled far enough to reach the zone.

Inside-zone interactions

Raw interactions recorded inside a configured zone provide the strongest coordinate-based evidence that the element was used.

If no valid zone is configured, inside-zone evidence is unavailable. It is not zero.

Sequential events and raw-zone evidence disagree

If step progression exists but raw inside-zone interactions are zero, Wheremore treats this as a measurement mismatch. It does not automatically claim that visitors failed to interact or that the zone is misaligned.

Use a controlled test to check:

the live element renders;
the selector or form mapping is correct;
the interaction creates the expected step event; and
completion creates the expected path-completion record.

Only adjust zone coordinates when the visual overlay clearly shows that the zone is materially outside the live element.

Form-submission tracking

For embedded or indirectly trackable forms, field-level coordinate clicks can be intentionally limited.

Form starts, submissions and path completions are then more authoritative than field-click evidence. Missing field clicks must not be presented as proof that visitors ignored the form.

Wheremore only refers to HubSpot when the collected evidence explicitly identifies a HubSpot form, script, iframe, redirect or submission event.

The analysis can use session-level conversion evidence in both modes.

With valid consent and a persistent visitor ID, durable conversion processing can recognise qualifying progress across visits in the same browser and device.

Cookieless tracking

Sessions remain separate. Wheremore can still analyse path entrants, step progression, form starts, completions and page behaviour, but it cannot assume that separate sessions belong to the same person.

Cookie identity does not provide reliable cross-device identity.

Add a recommendation to the Action Board

Select Add to Board on a supported step or journey recommendation.

If more than one eligible project is available, choose the project that should receive the task.

The Board item carries the source path, step or journey, evidence, priority, selected scope and intended outcome signal. Wheremore can then compare later results with the recorded baseline where enough evidence is available.

Adding a task does not prove that the recommendation caused a later improvement. Outcome measurement shows whether the selected signal changed after implementation.

Duplicate submissions of the same prepared recommendation are suppressed.

Ask Wheremore

Use Ask Wheremore about this CRO analysis to ask follow-up questions about the completed reports.

Useful questions include:

Which step should we test first?
Which conclusion is based on conversion truth rather than broader page behaviour?
What controlled test would confirm this measurement mismatch?
Which recommendation is suitable for the Action Board?
How does Cookie or Cookieless mode affect this conclusion?

Ask Wheremore uses the stored completed analysis. It does not replace missing evidence or make an incomplete job complete.

Recommended workflow

Start with all campaigns and a broad audience to understand the whole path.
Check path entrants, form starts and completed sessions.
Identify the largest source-of-truth leak.
Review the affected step report.
Confirm whether the evidence represents exposure, inside-zone interaction or accepted step progression.
Compare a campaign, audience or device only when the sample is useful.
Use Session Replay and Page Engagement for supporting evidence.
Run a controlled conversion test when measurement is uncertain.
Add one clear, measurable action to the Board.
Reanalyse after enough post-change data has accumulated.

Data timing and limitations

The analysis is not real time.
Recent interactions can appear before conversion and attribution processing is complete.
Screenshots can differ from the live visitor experience when content is delayed, personalised, consent-gated or embedded.
Zone exposure and page-level clicks do not prove form interaction.
Attribution explains touch context, not causation.
Small cohorts provide directional evidence.
Cookie tracking is not reliable cross-device identity.
A recommendation is a hypothesis to test, not proof of the cause of conversion or drop-off.

Troubleshooting

No conversion paths are available to analyse

Select a website.
Confirm the active workspace.
Confirm the website has an active conversion path.
Confirm the path belongs to the selected website.

No steps are found

Open Manage Conversions and confirm the path contains ordered steps.

A step analysis failed

Select Retry Step.
Confirm the step still belongs to the path.
Confirm its webpage and zone are available.
Check that screenshot capture and background processing are working.

Journey analysis failed

Every configured step needs one complete server-side step report. Retry failed steps, then select Retry Journey.

The report shows no progression

Confirm the selected period overlaps the path's active dates.
Confirm relevant pages are tracked.
Test the configured zones or form-submission mapping.
Allow conversion and funnel processing to complete.
Remove restrictive campaign, audience or device filters temporarily.

The report says zone evidence is unavailable

Check whether the relevant device has a valid zone. An intentionally unconfigured device is out of scope and does not automatically require a new zone.

The screenshot does not show the form

Check the live page. Embedded, consent-gated or delayed forms may not appear in capture even when they render for visitors. Use a controlled live test before diagnosing a UX failure.

Add to Board is disabled

The prepared Board evidence is no longer available. Rerun the analysis to create a fresh signed recommendation.

Recent changes are not reflected

Allow conversion, funnel, behaviour and attribution processing to finish, then rerun the analysis with a period containing sufficient post-change data.