Foundation OKR Writer: Brainshelf Resurface Q3
Sample: foundation-okr-writer. Brainshelf Resurface Q3 2026 OKR Set
Scenario
Brainshelf’s Resurface team shipped algo v1 in March 2026 and validated it with the ~500-user beta cohort (per the existing foundation-stakeholder-update brainshelf sample). Beta users showed 3.4x retention versus matched non-Resurface controls [fictional]. Algo v1 has now rolled out to all activated users.
For Q3 2026, the Resurface PM priya-pm needs OKRs that move from “we proved Resurface works for early adopters” to “Resurface is the habit-forming reason members come back each week.” The forcing function is a company-level OKR to increase 30-day member retention from 31% to 42% [fictional] by year-end.
The PM runs /okr-writer in Guided mode (default).
Source Notes:
- Brainshelf is fictional
- All metrics
[fictional] - Continuation of the brainshelf Resurface thread established in the
foundation-stakeholder-updateandmeasure-experiment-resultsbrainshelf samples - Hofmann’s “habit loop” model and Cutler’s North Star input-tree are the writer’s underlying retention-mechanism heuristics
Prompt
/okr-writer
Scope: Resurface team, BrainshelfCycle: Q3 2026 (July 1 to September 30)Level: teamOKR type: aspirational
Strategic context:- Company objective: increase 30-day member retention from 31% to 42% [fictional] by year-end 2026.- Resurface team owns the algorithm-driven content surfacing system.- Algo v1 shipped March 2026; beta cohort showed 3.4x retention lift vs control [fictional]; rolled out to all activated users in May.
Empowerment signal: empowered team. Leadership picks problems, team picksKRs and bets.
Draft objective: "Resurface should be the reason people come back."
Candidate KRs (rough):- More weekly engagement with Resurface- Resurface-engaged users retain better- Don't surface low-quality items
Constraints:- Engineering capacity: 3 engineers, 1 ML engineer, 1 designer- No new platform dependencies- Cannot influence top-of-funnel acquisition (other teams)
Source of truth: company OKR doc in Linear at L-OKR-Q3-ResurfaceOutput
# OKR Set: Brainshelf Resurface Team, Q3 2026
(Mode: Guided. Empowered-team signal: empowered. Disclosure section omitted.)
## Context
- Scope: Resurface team- Cycle: 2026-Q3 (July 1 to September 30, 2026)- Level: team- OKR type: aspirational- Empowered-team signal: empowered (leadership assigns problem; team picks KRs and bets). No Disclosure section needed.- Source of truth: Linear at L-OKR-Q3-Resurface. This artifact is a planning input, not the live tracker.- Strategic input: Company OKR to increase 30-day member retention from 31% to 42% [fictional] by year-end. Resurface is the largest single retention lever per H1 2026 cohort analysis (3.4x retention multiplier in the beta cohort vs control).- Assumptions: - The 3.4x multiplier observed in beta will hold or attenuate at general rollout. Treated as working hypothesis; KR2 design lets us measure the actual multiplier on the broader population. - Algo v1 quality is good enough to scale. KR3 guardrail catches degradation if rollout reveals issues the beta did not. - Push notification infrastructure exists and is approved for Resurface use.
## Objective
Make Resurface the reason members come back to Brainshelf each week.
(One qualitative, directional statement. Replaces "the reason peoplecome back" with language that signals desired member behavior at aspecific cadence.)
## Key Results
- KR1: Increase weekly Resurface-active members from 22% to 41% [fictional] of activated members by 2026-09-30. - Metric: percentage of activated members who interact with at least one Resurface-surfaced item in any 7-day window, averaged over the trailing 4 weeks. - Baseline: 22% [fictional] (June 2026 trailing-4-week average, post-rollout). - Target: 41% [fictional] by 2026-09-30. - Evidence source: Mixpanel `resurface_item_interaction` event, segmented by activated-member cohort. - Owner: priya-pm. - Indicator class: leading. - Confidence: medium. The 22 to 41 jump nearly doubles current state and assumes algo v1.5 plus push triggers deliver.
- KR2: Increase 30-day retention among Resurface-engaged members from 56% to 68% [fictional] (vs 31% baseline among non-Resurface members). - Metric: percentage of members who engaged with Resurface at least once in week 1 and returned at least once in days 8-30. - Baseline: 56% [fictional] (Q2 2026 cohort). - Target: 68% [fictional] by 2026-09-30. - Evidence source: cohort dashboard `resurface-30d-retention-q3`. - Owner: priya-pm. - Indicator class: lagging (30-day window). - Confidence: medium. Depends on whether the multiplier holds at scale.
- KR3 (guardrail): Hold member-reported relevance score at or above 4.2 / 5 [fictional] across all Resurface impressions. - Metric: median relevance score from in-product post-interaction survey ("How relevant was this to you?"). - Baseline: 4.2 / 5 [fictional] (Q2 2026 average). - Target: maintain or improve. - Evidence source: post-interaction survey (~5% sample rate). - Owner: priya-pm. - Indicator class: guardrail. - Confidence: high. Protects against the failure mode where push triggers boost engagement but degrade perceived quality.
## Initiatives as Bets
- Initiative 1: Algo v1.5 (early-August ship target). - Expected KR impact: KR1 primarily, KR2 secondarily. - Hypothesis: incremental tuning of the v1 algorithm based on beta findings will improve hit rate on first-impression items, increasing week-1 engagement which in turn drives retention. - Dependency: ML engineer plus 2 weeks of training data refresh.
- Initiative 2: Push notification triggers for high-confidence surfaces. - Expected KR impact: KR1 primarily. - Hypothesis: members who receive a push for a high-confidence Resurface item will engage at 2x the rate of in-app-only surfacing [fictional from beta data]. - Dependency: Notifications platform team. No new asks; existing infra supports this.
- Initiative 3: Personalized Resurface onboarding (first session). - Expected KR impact: KR2 primarily (early engagement predicts retention). - Hypothesis: members who experience Resurface in their first session have higher week-1 engagement and thus higher 30-day retention than members who discover it later. - Dependency: Onboarding team for first-session UX integration.
## Guardrails and Health Checks
- Guardrail 1: Member-reported "Resurface item felt repetitive" rate stays below 8% [fictional] threshold (Q2 baseline 5%). - Why: algo v1.5 may over-surface a narrow content set. If users perceive repetition, engagement gains are hollow.
- Guardrail 2: Push notification opt-out rate stays below 6% [fictional] (Q2 baseline 3.1%). - Why: aggressive push triggers risk users disabling all Brainshelf notifications, which damages retention well beyond Resurface.
## Alignment Notes
- Parent or strategy link: Company OKR "Increase 30-day member retention from 31% to 42% [fictional] by year-end." This Resurface OKR is the primary retention lever; Onboarding team and Sharing team have secondary retention OKRs.- Peer dependencies: - Onboarding team owns Initiative 3 first-session UX. Confirmed capacity in Q3 planning. - Notifications platform team owns push infrastructure for Initiative 2. No new asks from us this quarter.- Known conflicts: - The Sharing team's Q3 OKR includes "Increase items shared per week by 30%." If they succeed, the surface-able content pool grows faster than algo v1.5 can adapt. Mitigation: weekly sync between teams to align retraining cadence with sharing-volume changes.- Out of scope this cycle: - Monetization (no team owns; not Resurface's lever). - Net-new member acquisition (Growth team). - Resurface for non-activated members (deferred until activation flow is settled).
## Quality Audit
- Strategic fit: pass. Clear link to company retention OKR via "Resurface is the largest single retention lever" logic.- Objective quality: pass. "Reason members come back each week" implies a habit-forming behavior change, not a feature delivery.- KR outcome quality: pass. All three KRs measure member behavior or perceived-quality outcomes; no "ship algo v1.5" KRs.- Measurement quality: pass. All three KRs cite specific Mixpanel events or in-product survey instruments.- Product influence: pass. Team controls all three KRs' levers via algo, push triggers, and onboarding UX.- Focus: pass. One Objective, three KRs (within 2 to 4 range).- Guardrails: pass. KR3 plus two health metrics protect against the predictable failure modes (perceived repetition, push fatigue).- Alignment: pass. Parent OKR named, peer dependencies named, conflict with Sharing team flagged with mitigation.- Operating rhythm: pass. Quarterly cycle; weekly check-ins assumed (default for product team aspirational OKR).- Integrity: pass. Team-level OKR, no compensation coupling, no fabricated data.- Empowered-team Disclosure: not-applicable. Empowered signal; section omitted.
## Open Questions
1. The 3.4x retention multiplier was observed in a 500-user beta cohort that self-selected into the program. Will it hold at population scale, where members did not opt in? KR2 design will measure this, but the team should pre-commit to a "validated multiplier" threshold below which the strategy needs to change next cycle.
2. The push notification opt-out guardrail (Guardrail 2) is an aggregate metric; if a sub-segment of users opts out at high rates, the median may hide it. Recommend instrumenting opt-out by user segment (engagement tier, account age) for Q4 visibility, even if not graded this cycle.
3. Algo v1.5 is described as "incremental tuning." If the team finds v1.5 is materially the same as v1, should they spike on a v2.0 rewrite mid-cycle? Recommend pre-deciding the threshold (KR1 trending below X% by week 6) that would trigger that pivot.
## Suggested Next Step
Run `/okr-writer` again in Audit Only mode after the cross-team Q3planning workshop. Specifically check that the Sharing team's KRconfirmation does not invalidate KR1's assumption that surface-ablecontent pool stays roughly stable. Also confirm Open Question 1'sthreshold with company leadership before cycle start.