Growth Engineering Playbook · Wave 2
What should we show next — and what should we never show?
A governed product recommendation engine for the fictional Northstar Outfitters catalog. Legible strategies — complementary, bought-together, upsell, similar, trending, same-brand — are blended by a merchandising objective, then filtered by business guardrails: in-stock only, a margin floor, return-risk suppression (fed by the support miner), and category diversity. Every recommendation explains which rules fired and on what evidence. Recommendations are a rules-and-governance problem, not a black box. Deterministic, client-side, catalog + behaviour signals only.
Loading catalog & behaviour data…
Recommendation Slate
Ranked by blended score. Click a recommendation for its full reasoning.
Why This Recommendation
Which strategies fired, their weighted contribution, and the evidence.
Strategy Contribution
How the objective weights each strategy, and how many made the slate.
Guardrails & Coverage
The business rules governing the slate — and catalog-wide coverage.