Score whether new customers feel taken care of.
Grades the first 90 days on the moments that make someone feel expected, not just whether they activated, and names the exact spot the welcome goes cold.
In plain English: “Is our onboarding actually any good?”
Inside: The moment-by-moment rubric, the cold-spot finder, the silent-stall check, and the fixes.
Install it in one line, or paste it in.
~/.claude/skills/ and runs automatically when it is relevant.Connect your context. Set it to your motion.
it sees where real customers stall versus where the playbook assumes they sail through.
it reads the human touches that actually happened, the calls, the check-ins, the silences.
it hears the friction in the customer's own words.
This was built to be lifecycle-agnostic. Set these to your motion:
The rubric is yours. The point is to grade the experience your customers should feel, not a generic checklist someone else outgrew.
| Set this | What it is | Default / Example |
|---|---|---|
| LIFECYCLE | your onboarding window | first 30 / 60 / 90 days |
| MOMENTS | the touchpoints you grade | welcomekickofffirst valuehabitcheck-in |
| VALUE_EVENT | the moment they first get value | your product's aha moment |
| HUMAN_TOUCHES | where a person shows up | kickoff callcheck-inQBR |
| SIGNALS | how you see a customer stalling | login gapsetup abandonsupport spike |
| VOICE | the feeling you want to leave | your brand's version of "taken care of" |
Everything the skill does, in full.
Grades the experience a new customer actually has, not just whether a checkbox got ticked. Most onboarding scoring measures activation: did they log in, did they finish setup. This scores the feeling: were they expected, were they guided, did anyone notice when they went quiet. It walks the first 90 days moment by moment, scores each one, and hands back the specific places the welcome went cold and what to do about it. In commoditized software the experience is the moat, so this measures the moat.
- 1Score the moments, not the checkboxes
Walk the first 90 days as the customer lives it: the welcome, the kickoff, the first real value, the habit forming, the first check-in. Each moment gets a score on two axes: did it happen, and did it feel like someone was expecting them. A completed setup with no human warmth scores lower than it looks.
- 2Find where the welcome goes cold
Name the exact moment the attention drops off. For most teams it is right after setup: the sale got all the energy, activation got a sequence, and then the customer is alone. That cold spot is where early churn is born, and it has a location you can point to.
- 3Grade the human touches
Automation onboards; people make customers feel chosen. Score where a human actually shows up versus where the customer is handed to a drip. A high-value customer getting a templated welcome is a scored miss, not a nice-to-have.
- 4Catch the silent stall
Score whether anyone notices when a new customer goes quiet. The worst onboarding failure is invisible: they stopped, and no one reached out, because no signal was wired to a human. If the stall is silent, that is the finding.
- 5Turn the score into fixes
Every low-scoring moment ends in a specific move: the touch to add, the signal to wire, the human to insert. Not "improve onboarding." The exact cold spot and the exact fix.
- Every graded moment scores both dimensions: did it happen, and did it feel intentional.
- The cold spot is named with a location, never "onboarding could be better."
- Human touches are scored separately from automation, never blended into one number.
- The silent-stall check is explicit: is a quiet customer noticed, yes or no.
- Every low score ends in a specific fix, never a generic recommendation.
ONBOARDING AUDIT · [your product] · first 90 days Moment Happened Felt intentional Score Welcome yes templated 2/5 Kickoff call yes warm, tailored 5/5 First value partial no guidance 2/5 Habit forming no - 1/5 Day-30 check-in no - 0/5 COLD SPOT: right after first value. The kickoff is great, then the customer is alone. No habit nudge, no Day-30 human touch. SILENT STALL: not caught. No signal wired to a human when a new account goes quiet in week 3. FIXES: 1. Wire a week-3 inactivity signal to the CSM, not a drip. 2. Add a habit nudge tied to the aha moment, day 7 to 14. 3. Make the Day-30 check-in a human note, not an NPS survey. Onboarding experience score: 2.0 / 5 (strong start, cold finish)
The moments, the window, and the scoring bar are yours to set. The examples here suited a 90-day mid-market onboarding. A self-serve motion compresses the window; an enterprise motion stretches it. The logic, score the feeling not the checkbox, find the cold spot, catch the silent stall, does not change.
Where an operator takes this next.
The audit is step one. Here is where an operator takes it once the score is real.
You grade the welcome once, then keep it warm on purpose.
Have a scheduled Claude task read Amplitude or your warehouse nightly and DM the CSM in Slack the new accounts that went quiet this week, so no welcome ever goes cold unnoticed.
Re-run the audit each month against the latest cohort so a regression in the experience shows up as a dropping score, not as churn three months later.
Hand the findings to an offboarding audit so you can see the full arc: how you greet them and how you say goodbye, and whether the two match.
One skill is the on-ramp.
A single skill does one job. Chained into a playbook, or run as a full build, it becomes a system. Here is where this one plugs in.