| name | getting-started-audit |
| description | Measure and shorten time-to-first-success — walk the getting-started path from a clean state, time it, rank friction points by where developers drop off, and make a fix-or-document call on each. Use whenever you need to know how good (or bad) onboarding really is. |
Skill: Getting-started audit
Measure the one number that decides adoption — time-to-first-success — and find the
friction that's inflating it. Most developers decide in the first ten minutes.
When to use
- You don't know how long it takes a new developer to get a real result.
- Sign-ups are healthy but activation (first app / retention) is weak.
- Before investing in reach — reach that lands a dev on a broken first-run is waste.
Procedure
- Start from a genuinely clean state — new account, no cached creds, no
insider knowledge. Anything you "already know" is a friction point a real
developer will hit.
- Time it to first success. Record the wall-clock minutes from signup to the
first real result (first successful call, first working app). That's TTFS.
- Log every friction point — each place you paused, guessed, hit an error, or
left the page. Note where in the funnel (signup → first call → first app) it
sits, because friction earlier costs more.
- Rank by drop-off impact, not by how annoying it felt to you. Friction at
"signup → first call" loses more developers than friction deep in an advanced
flow.
- Make a fix-or-document call per friction point (tree 2 in
../../knowledge/devrel-engagement-decision-trees.md):
product-pain → product-feedback ticket; sound-but-undiscoverable → content task.
- Capture it in the
getting-started-audit template.
Output
A measured TTFS, a ranked friction list, and a fix-or-document disposition per
item — the input to both the content plan and the product-feedback brief.
Anti-patterns (the hook flags these)
- A getting-started doc with no explicit first-success milestone.
- "Improving onboarding" with no TTFS measurement to move.
- Papering over a product flaw with a longer tutorial instead of filing the bug.