Example: turning a workflow into metrics
This example follows the same registration workflow from Section 2 and turns it into events, metrics, dimensions, and decisions.
The point is not to create the perfect registration dashboard. The point is to show the reasoning from behaviour to evidence to measures.
Starting workflow
Workflow:
Register for an account
User intent:
Create an account so I can access the product.
Entry point:
User starts account creation.
Steps:
1. View registration form
2. Enter details
3. Submit form
4. Verify email
5. Access product
Completion:
User accesses the product for the first time.
This workflow gives the team a behavioural structure. Now the team can decide what evidence would make the steps observable.
Event design
| Workflow step | Event | What the event proves |
|---|---|---|
| View registration form | registration.form_viewed |
The user reached the form |
| Submit form | registration.form_submitted |
The user submitted registration details |
| Validation error shown | registration.form_error_shown |
The user encountered a form error |
| Verify email | registration.email_verified |
The user completed email verification |
| Access product | registration.completed |
The user reached the product after registration |
The events do not explain everything, but they provide the behavioural evidence needed to create metrics.
Metric definitions
| Product question | Metric | Example definition |
|---|---|---|
| Are users starting registration? | Registration start count | Count of users who trigger registration.form_viewed |
| Are users able to submit the form? | Form submission rate | Users who trigger registration.form_submitted ÷ users who trigger registration.form_viewed |
| Are users encountering form problems? | Form error rate | Users who trigger registration.form_error_shown ÷ users who trigger registration.form_viewed |
| Are users completing verification? | Email verification rate | Users who trigger registration.email_verified ÷ users who trigger registration.form_submitted |
| Are users reaching the product? | Registration completion rate | Users who trigger registration.completed ÷ users who trigger registration.form_viewed |
Each metric answers a different question. Together they help the team understand where the workflow is healthy and where it may need investigation.
Useful dimensions
The team should not only look at the overall metrics. It may need to compare behaviour by:
- entry point
- device type
- user type
- workflow version
- channel
- error type
For example, a stable overall completion rate could hide a mobile problem. A high form submission rate could still hide a specific validation error. A lower completion rate for invited users could point to the invitation flow rather than the form itself.
Dimensions turn a broad signal into a more focused investigation.
Decisions the metrics can support
The metrics should support practical decisions:
| Signal | Possible decision or investigation |
|---|---|
| Low form submission rate | Review field design, guidance, trust signals, or perceived effort |
| High form error rate | Improve validation rules, labels, hints, or error messages |
| Low email verification rate | Investigate confirmation screen, email content, deliverability, or resend behaviour |
| Low first product access | Improve post-verification handover or first-use experience |
| Mobile underperformance | Prioritise mobile usability review or performance checks |
The metric does not decide for the team. It helps the team ask the next better question.
What this example shows
The workflow creates the structure. The events provide evidence. The metrics summarise patterns. The dimensions add context. The decisions explain why the measurement matters.
The important move is not the specific registration example. It is the reasoning: do not invent the metric in isolation. Trace it back to behaviour and forward to use.
Key takeaway
A metric is strongest when it can be traced back to a workflow step and forward to a decision.
That traceability is what turns event data into product measurement.