Implementation worksheet · 6 min read
An onboarding flow teardown rubric for product teams
Walk the flow as a new user on a fresh account, record every screen, then score eight criteria from 0 to 2 with one line of evidence each: time to the first value event, steps before value, whether questions asked change anything, clarity of the next action, skip and resume, keyboard and screen reader basics, handling of errors and empty states, and whether guidance points at real product actions or at the interface itself. A score without evidence doesn't count. For a competitor's flow, score only what you observed, record the date, and don't guess at their metrics.
Teardowns are popular and often subjective: screenshots with commentary and a verdict shaped by visual taste. A rubric makes two reviewers land in the same place, or shows exactly where they disagree. The actor is a product team reviewing its own onboarding before a redesign, or studying another product's. Boundary: from signup or first launch to the first value event, on one platform, as observed on a stated date.
Put it into practice
1. Fix the persona and the goal first
Which kind of user, trying to do which job, reaching which value event. A flow that's excellent for an admin setting up a workspace can be poor for an invited member, and scoring both at once produces mush.
2. Record the run on a clean account
Screen recording, timestamps, a fresh account and a clean browser profile. Note the date and any version shown, because flows change often and an undated teardown goes stale without anyone noticing.
3. Score each criterion 0, 1 or 2 with evidence
0 means missing or blocking, 1 means present with friction, 2 means clear. Every score carries a timestamp, a screenshot or a quoted line of copy. 'Feels cluttered' isn't evidence; 'three competing prompts on the first screen' is.
4. Have two people score independently
Then compare. Where scores differ by two points, the disagreement is itself a finding: often the criterion is ambiguous, or the reviewers imagined different personas.
5. Keep observation and inference apart
For another company's flow you can see the steps, the copy and the timing. You can't see their activation rate, and 'this probably converts well' is a guess dressed as a finding. Leave it out.
6. Turn each low score into one specific change
Every 0 or 1 becomes a proposal tied to its criterion, such as 'move billing after the first project', not 'simplify onboarding'. Rank proposals by which criterion sits closest to the value event. In the illustrative scoring below the flow totals 8 of 16, and its two zeros become the first two proposals.
7. Re-run after changes with the same persona
Same account type, same goal, new date. The rubric is most useful as a before-and-after record of your own flow, where you can also check the change against real metrics.
Teardown rubric
Copy this structure into your review document and record your observed result for each row.
| Criterion | 2 means | 0 means | Illustrative score and evidence |
|---|---|---|---|
| Time to first value event | Reached in one sitting without help | Not reached in the session | 1: 14 minutes; import option hard to find |
| Steps before value | Only steps the value event needs | Unrelated setup comes first | 1: billing details requested before the first project |
| Questions asked | Each answer changes what follows | Answers collected and never used | 0: role question changes nothing |
| Next action clarity | One obvious next step per screen | Several competing prompts | 2: single primary button throughout |
| Skip and resume | Skippable; resumes where the user left | Restarts from the beginning | 1: skip works; resume restarts the tour |
| Keyboard and screen reader basics | Operable and announced | Traps focus or unlabeled controls | 1: focus lost after the tour closes |
| Errors and empty states | Explain and offer a way forward | Blank screen or generic failure | 2: empty dashboard offers sample data |
| Guidance points at real actions | Tips lead the user into doing the task | A tour of the interface with no task | 0: 7-step interface tour before any task |
A failure worth checking
Scoring what the reviewer likes. Two reviewers tear down the same flow: one scores it highly for polished animation, the other poorly for length. Without criteria tied to the user's goal, the teardown turns into a taste debate and the redesign follows whoever is more senior. The rubric doesn't remove judgement; it makes each judgement point at evidence someone else can check.
Common questions
Can we publish a teardown of a competitor's onboarding?
Describe what you observed on a stated date, use only screenshots you're entitled to use, and make no claims about their results. Check with counsel on screenshot and trademark use if the teardown will be public. A dated, observation-only teardown stays fair; an undated one ages into something misleading.
Does a rubric replace usability testing?
No. Reviewers already know the domain and read an interface differently from a real newcomer. Use the rubric to find the weak stretches, then watch new users go through those stretches before redesigning them.
Basis and scope
This is a proposed implementation method using illustrative examples, not a measured benchmark or a customer case study. Prepared with AI assistance. Validate product-specific behavior against current documentation and your own test environment.