💡 Queen performance intelligence with longitudinal scoring and predictive requeening guidance

[github-actions[bot]]

Summary

Aggregate multiple inspection observations — brood pattern quality, temperament assessments, laying rate indicators, presence of supersedure or swarm cells, and queen age — into a continuous queen performance score that trends over time. When the score declines below colony-specific thresholds, surface requeening guidance with timing recommendations, method options, and seasonal context. No competitor offers structured queen evaluation scoring.

Market Signal

Queen failure is the single largest cause of colony loss for hobbyist beekeepers. Research from the Urban Beekeeping Hub (2026) shows a 73% first-year success rate for mentored beekeepers vs 34% without mentoring — structured evaluation guidance (the kind a mentor provides) dramatically improves outcomes. Current beekeeping apps (HiveTracks, HiveSense, Beentry, HiveBook) all capture queen observations as unstructured notes but none synthesize them into a longitudinal performance score or predictive requeening advisory. HiveTracks v3.0 added AI plant identification but no queen intelligence features.

User Signal

The PRD's hive lifecycle management and the recommendation engine's action + rationale + confidence contract provide the foundation. The skill progression system targets exactly the pattern-recognition gap that beginners have with queen evaluation. The PRD user journey for Hannah (anxious → guided → confident) maps directly to queen performance scoring reducing decision uncertainty at the moment it matters most.

Technical Opportunity

Queen observations are already captured in the inspection schema (architecture.md defines inspections domain with structured logging). A scoring algorithm aggregates existing data fields — no new data capture needed. The recommendation engine can consume the score as an input signal, generating contextualized requeening guidance. The pgvector-based embedding store enables similarity matching against known queen performance patterns from the analytical data layer. Queen lineage tracking (Discussion #454) provides the genealogical context.

Assessment

Dimension	Score	Rationale
Feasibility	high	Algorithmic scoring from existing inspection data fields. No new hardware, sensors, or AI models required. Leverages existing recommendation engine contracts.
Impact	high	Queen failure is the #1 colony loss driver. Structured scoring replicates the mentor's pattern recognition that produces 73% vs 34% success rate improvement.
Urgency	med	No competitor currently offers this, so there's no competitive pressure to rush — but the mentoring-equivalent value positions Broodly uniquely for the beginner segment.

Adversarial Review

Strongest objection: Queen evaluation is highly subjective and varies by breed, season, and beekeeper experience. A simplified score could mislead beginners into requeening productive queens or ignoring issues in queens with temporarily low scores.

Rebuttal: The score is not a replacement for expert judgment — it's a structured way to surface patterns that beginners miss. Breed and seasonal normalization calibrate expectations. The recovery-oriented tone ("She's showing some signs of declining brood pattern — watch for X next time, and here's what to do if it continues") prevents panic while maintaining awareness. For experienced beekeepers managing 20+ colonies, the score is a time-saving triage tool across many queens.

Suggested Next Step

Define the queen performance scoring algorithm with input weights for each observation type (brood pattern: 30%, laying rate: 25%, temperament: 15%, supersedure signs: 20%, age factor: 10%). Validate the scoring model against published queen evaluation rubrics (e.g., Bee Informed Partnership's queen scoring criteria). Create a UX design story for the queen performance trend visualization within the hive health card.