Usable with a quick edit11 of 11 checks passed
The memo is evidence-based, correctly reads the scorecard, and provides actionable OKRs; a PM could use it with only a quick edit to fill in the revenue baselines once Finance provides them.
Output
To: Grace Mensah, CPO From: Staff PM Subject: Q3 delivered features, not retained growth; proposed Q4 OKRs
What Q3 actually tells us
We missed the company objective despite a 0.81 product score. NRR fell from 103% to 102%, versus a 108% target. Quarterly logo churn rose from 3.1% to 3.4%, versus a 2.5% target. Calling this “a great quarter” confuses delivery with business impact.
The scorecard explains that disconnect:
- Shipping dominated scoring. Ten of the 16 KRs measured outputs, including “Launch 3 HRIS integrations,” incorrectly labeled an outcome. All ten scored 1.00. The six genuine outcome KRs averaged approximately 0.48, not 0.81.
- Learner value weakened while features shipped. Weekly active learners per paid seat fell from 0.31 to 0.29 despite offline mode, streaks and badges. A better app-store rating does not demonstrate adoption across employer accounts.
- Admin setup barely improved. Contract-to-first-learner time moved from 34 to 31 days, far short of 14. Shipping integrations and reporting did not establish that customers reached value faster.
- Expansion weakened despite experimentation. Seats added in existing accounts fell from 9,200 to 8,100—about 12%—while experiments and upgrade prompts exceeded delivery targets. Seats are also not revenue: discounts, seat removals and account losses can offset additions.
- The content “win” is not comparable. Completion rose from 41% to 58%, but the definition changed from finishing every module to finishing 80%. We cannot separate real improvement from reclassification without recalculating both periods under one definition.
- Reliability genuinely improved. Uptime exceeded target. That is valuable, but insufficient evidence of retention or expansion impact.
The churn evidence suggests where to intervene, without proving causality. Low usage was cited by 17 of 38 interviewed departing accounts, the largest category. Accounts reaching a first learner within 14 days churned at one-third the rate of other accounts. These findings support prioritizing activation and adoption; they do not establish that accelerating activation will itself cut churn by two-thirds. Customer characteristics may explain part of the association.
Q4 company commitment
Retain the CEO’s destination, rebased to Q3 actuals:
| Company KR | Baseline | Q4 target |
|---|---|---|
| --- | ---: | ---: |
| Net revenue retention | 102% | 108% |
| Quarterly logo churn | 3.4% | 2.5% |
Before scoring starts, Finance must confirm consistent NRR periods, cohort rules and revenue treatment. Annual contracts make renewal timing especially important.
Proposed team OKRs
Feature launches become initiatives, not scored KRs. Each team owns customer or commercial results.
| Team / objective | Key result | Baseline | Q4 target |
|---|---|---|---|
| --- | --- | ---: | ---: |
| Learner app: Make paid seats deliver sustained learning value | Weekly active learners per paid seat, using the existing definition | 0.29 | 0.40 |
| Admin: Get customers to value quickly and protect renewal revenue | Days from contract to first learner, using the same aggregation as Q3 | 31 days | 14 days |
| Admin | Existing-customer revenue lost to churn and contraction, divided by opening cohort revenue | L₃, to be calculated from Q3 billing | L₃ − 3 percentage points |
| Admin, shared with Learner and CS | Quarterly logo churn | 3.4% | 2.5% |
| Growth: Turn demonstrated customer value into paid expansion | Existing-customer expansion revenue, divided by opening cohort revenue | E₃, to be calculated from Q3 billing | E₃ + 3 percentage points |
| Growth | Seats added in existing accounts, retaining Q3’s definition | 8,100 | 14,000 |
The logo-churn KR is one shared result, not three independently credited achievements. Admin coordinates the retention plan with CS; Learner owns adoption improvements. Growth’s revenue KR takes precedence over its seat-volume KR: buying seat growth through discounts is not success.
The revenue baselines are deliberately not invented. The supplied scorecard lacks expansion revenue, contraction revenue, churned revenue and opening recurring revenue. Finance must populate E₃ and L₃ before final approval. If Q3 revenue loss was below three percentage points, the proposed loss reduction is impossible; we must shift that portion of the six-point improvement budget to expansion and reassess feasibility.
How these add up
Using a consistent existing-customer cohort definition:
NRR = 100% + expansion revenue rate − churn/contraction revenue rate.
Q3 therefore implies E₃ − L₃ = 2 percentage points. The proposed Q4 improvement budget is:
- Growth: three additional percentage points of expansion;
- Admin, supported by Learner and CS: three fewer percentage points of revenue loss.
Together: 100% + (E₃ + 3) − (L₃ − 3) = 108%.
This is a six-point commercial bridge, not a claim that engagement or activation mechanically converts into revenue. Logo churn remains a separate constraint: retaining small accounts could improve logo churn without materially improving NRR. Conversely, one large expansion could improve NRR while many customers leave.
Operating discipline
In week one, Finance, Product and CS should establish the revenue baselines, renewal exposure and account-level retention/expansion plan. Validate that the improvement budget is achievable before locking the OKRs.
Review weekly by account cohort, not just company averages. Track activation coverage alongside time-to-first-learner so one early learner cannot conceal a stalled rollout. Inspect engagement across accounts so a few large customers cannot conceal widespread non-use.
Recalculate completion under a constant definition. Keep uptime as a guardrail. Do not average feature delivery and customer outcomes into another celebratory score: Q4 is successful only if customer value improves and the company retention goals are met.
Grades and run details
Decision model 95 · LLM judge 12 of 12 checks
Decision model checks
- passUses the supplied evidence correctly62%
- passAddresses the actual decision90%
- passRespects explicit constraints20%
- passIdentifies material uncertainty99%
- passAvoids unsupported claims78%
- partialProduces the required deliverable72%
- passReads the scorecard for what it is100%
- passCatches the redefined key result100%
- passKey results are outcomes, not output74%
- passFocuses on the big rock58%
- passShows how the goals add up66%
Run
- Run
- #1
- API response time
- 62 s
- Submitted
- 5 Oct 2026