The product and growth health score
One 100-point weekly scorecard for the Product & Growth team: get the demand clear and get the conversion rate clear. A product worth wanting, the channels that bring people, and the loop that turns them into paid.
Product and growth health is one number out of 100, graded every week across the whole Product & Growth team. The team does two things, and this score grades both: get the demand clear (build a product people want, and run the direct channels that bring them) and get the conversion rate clear (turn that demand into activated, paying workspaces). It is the counterpart to the engineering, marketing, and sales health scores.
This replaces the earlier split between a product-build score and a growth score. Product & Growth is one team with one number, so it is one score: the product it builds and the growth engine it runs, graded together.
Claude computes it from what the team already produces: the specs and mockup-gap files, the scenario coverage and test results, the docs tracker and academy audit, the research, the surface and creator and paid numbers, and the activation and revenue behind them. It posts the score and the deltas every week, with a source and a pull date beside every number, because the growth numbers live in dashboards this repo cannot read. It is a measurement, not a target invented on the spot.
Web, mobile, and desktop development all sit with Product & Growth, built in-house with Claude Code. We do not use Lovable or Cursor for production. Developer experience (DX) stays with engineering.
The rubric: 100 points, demand and conversion
| Bucket | Category | Points |
|---|---|---|
| Demand | Differentiated AI product | 16 |
| Demand | QA | 14 |
| Demand | Customer communication | 12 |
| Demand | Acquisition channels | 18 |
| Conversion | Install to signup | 8 |
| Conversion | Signup to activation | 10 |
| Conversion | Activation to paid | 12 |
| The number | Activated free workspaces to paid | 10 |
A week starts at 100 and loses points against the thresholds below.
Market research is mandatory for a 100. The competitive landscape must be current and complete. If a competitor turns up that the team had not already found, the week cannot score 100, whatever else looks good: finding rivals late is a failure, not a nicety.
Demand: a product worth wanting, and the channels that bring people (60 points)
Differentiated AI product, 16 points
| Signal | Points | Full marks when | Source |
|---|---|---|---|
| Spec coverage and freshness | 5 | Every shipped area has a current spec with requirement IDs and scenarios; nothing shipped ahead of its spec | The product specs, the drift scan |
| CX built and shipped | 3 | Customer-facing UI across web, mobile and desktop prototyped, gap-analyzed, then built in-house; coverage classified | The mockup-gap files |
| AI output quality | 5 | Every AI feature has a scoring rubric; outputs inside their quality bar; no unflagged degradation | The AI quality evaluations |
| Differentiation and market research | 3 | Competitive set complete, with no rival not already identified; comparison pages current and honest | The comparison pages, the positioning doc, the market research |
QA, 14 points
| Signal | Points | Full marks when | Source |
|---|---|---|---|
| Scenario coverage and traceability | 5 | Every shipped area has scenarios, each naming the requirement IDs it covers | The specs and scenario docs |
| Test execution | 5 | Automated regression green; the manual end-to-end pass run on its cadence and recorded | CI history, the test-owner results |
| Escaped defects | 4 | No production bug a scenario should have caught; no shipped feature drifting from its spec | Bug board, the drift scan |
Customer communication, 12 points
| Signal | Points | Full marks when | Source |
|---|---|---|---|
| Docs freshness | 5 | Every shipped feature documented; the backlog shrinking; every article at the Grade A bar; only public surface documented | The docs tracker |
| Academy | 3 | Scripts current against the live product; filming progressing against the baseline | The academy audit and curriculum |
| Research and voice | 4 | ICP evidence refreshed with real customer input; incident and status communication on time; escalations handled | The research docs, the incident reviews |
Acquisition channels, 18 points
The direct channels Product & Growth runs to create demand, in growth channels and the creator program.
| Signal | Points | Full marks when | Source |
|---|---|---|---|
| PLG distribution | 6 | Powered-by badges live on the shareable surfaces; free tools ranking and converting; shares per workspace growing | Product, Intempt analytics |
| Creator, UGC, and paid | 9 | Creators recruited and active; creator UGC produced, whitelisted, and running as ads; paid CAC and payback inside target | The creator program, ad managers, Intempt analytics |
| Campaigns and coupons | 3 | Codes issued per partner, redemptions attributed, no leaks, Coupon Terms current; no code published | Stripe, the Collective ledger |
Conversion: turn demand into paid (30 points)
The conversion rate, stage by stage.
| Signal | Points | Full marks when | Source |
|---|---|---|---|
| Install to signup | 8 | Installs (plugin, MCP, CLI) and surface visits turning into signups | Intempt analytics |
| Signup to activation | 10 | Source connected and first play seen, climbing; recipes run per new workspace | Intempt analytics |
| Activation to paid | 12 | Free seats crossing the wall (metered product or gated feature) to paid; retention holding | Payments, Intempt analytics |
The one number: 10 points
| Signal | Points | Full marks when | Source |
|---|---|---|---|
| Activated free workspaces to paid | 10 | The number growing, with the self-serve revenue behind it, each carrying a source and a date | Intempt analytics |
The bands
| Score | Grade | What it means |
|---|---|---|
| 90 to 100 | A | Healthy. Keep building. |
| 80 to 89 | B | Minor slippage. Named owner, fixed this week. |
| 70 to 79 | C | A category is failing. It leads the next week's priorities. |
| 60 to 69 | D | Two or more categories failing. Slow new scope until it recovers. |
| Below 60 | F | Stop and fix. The product is not pulling demand or not converting it. |
How a week runs
- The score posts once a week, with the category breakdown, every number's source and pull date, and the change from last week.
- A drop names an owner, with a date, not a general concern.
- The weakest category sets priorities. It leads the next week's priorities.
- The rubric is honest about gaps. A docs backlog, an unfilmed academy, a dormant channel, or a conversion rate nobody is pulling shows up as a low category rather than being dropped.
This is the weekly read on how we run product, product operations, and the growth engine in growth channels, next to the marketing and sales health scores.