The client already had plenty of reporting. What was missing was a decision system — something that could take SKU, channel, season, and inventory signals and turn them into price moves a merchant would actually trust. We built that, and the workflow around it.
The business was optimizing inside disconnected spreadsheets, each with its own private assumptions about elasticity, discount depth, and what substitutes for what.
Recommendations had to respect merchant judgment, brand constraints, inventory position, channel conflict, and regional demand — at the same time, without picking favorites.
The model could not be a black box. Merchants needed to see why a move was recommended, what would break if assumptions shifted, and where they still owned the final call.
Demand response layer
We modeled demand sensitivity across product families, channel behavior, seasonality, and promo depth, then exposed it as a reusable layer the team could call into — not a one-off analysis that ages out.
Optimization harness
A scenario optimizer searched price and markdown moves against margin and inventory targets, while honoring the business rules merchants already argued about in planning meetings.
Merchant-facing review loop
The output was built for operators: the recommended move, the expected tradeoff, a confidence band, the constraint that bound, and the exact assumption a merchant could override if they disagreed.
- Started with a narrow category where the team had enough history to actually test elasticity assumptions.
- Compared recommendations against the past year of decisions before asking merchants to act on a live one.
- Once the review loop held up, expanded out of price and markdown into promotion planning, wholesale pricing, and seasonal campaign support.
Net margin lifted roughly five percent while demand still grew about ten.
Planning meetings stopped being subjective discount debates and started being scenario comparisons with the tradeoffs on the table.
The model survived past the first quarter because overrides and approvals were built into the workflow from day one, instead of bolted on after merchants pushed back.
Pricing AI sticks when it becomes a decision system, not just a model. The numbers matter. Adoption comes from making each recommendation legible to the merchant who has to defend it.
