Bỏ qua để đến nội dung

Entitlement-aware recommendation composition and lock-teaser policy

DomainsDOL EnglishProduct275 words1 min read
confirmedbyProduct Design

DEC-0077 - Entitlement-aware recommendation composition and lock-teaser policy

Phần tiêu đề “DEC-0077 - Entitlement-aware recommendation composition and lock-teaser policy”

Recommendation quality improved with freshness/diversification, but users can still receive too many locked items, creating friction right after return and reducing immediate learning momentum.

Recommendation set composition now follows entitlement-first actionability:

  • Available-now priority:
    • each set should prioritize items that user can start immediately under current entitlement.
  • Lock teaser cap:
    • at most 1 locked item per set when available inventory is sufficient.
    • locked teaser is placed at end of set.
    • teaser must show:
      • lock reason,
      • minimum eligible plan (Pro or Pro Max).
  • Low inventory fallback:
    • first relax by nearest-ladder within same program/assessment-form/skill scope,
    • only when still insufficient, allow additional locked items with short “limited available items” notice.
  • Seed compatibility:
    • if set uses reactivation seed handoff, seeded quick-win must be available-now.
  • Non-coercive boundary:
    • recommendation browsing never hard-gates user to upgrade,
    • upgrade remains optional via locked teaser interaction.
  • Keeps user in learning flow with minimal friction.
  • Preserves upgrade discovery without making recommendation feel like paywall feed.
  • Works with deterministic rules and scales easily.

Users return to learn first. Recommendation must maximize immediate start success before monetization prompts.

  • Product/UX impact:
    • higher first-attempt start likelihood from recommendation surfaces.
  • Data/logic impact:
    • recommendation payload needs available_now and minimum_eligible_plan fields.
  • Operational impact:
    • no extra operations overhead; rule-based composer only.
  • Option A: allow mixed locked/open items with no cap.
  • Option B: hide all locked items completely (no teaser).
  • No blocker for baseline lock.