Demand and pricing
Demand planning and pricing, under real uncertainty
Forecasting, markdown and price decisions built as one system. Probabilistic where it matters, constrained by how the business actually operates, and deployed on the cadence planning already runs on.
What this usually looks like
- The forecast is a single number, and nobody plans inventory against a single number.
- Markdown runs on rules written years ago that nobody wants to be the one to change.
- Price elasticity was estimated once, on data that could not identify it.
- Planners override the model every cycle, and no one has measured whether they are right.
What you get
Forecasts with uncertainty attached
Hierarchical where the data is sparse, probabilistic where the decision is asymmetric. Sized to the SKU and horizon the business actually plans on.
Decisions, not predictions
Markdown depth and timing, buy quantities, price moves. The output is what to do, with the forecast underneath it.
Constraints modelled honestly
Lifecycle limits, inventory positions, monotonic price response, contracted minimums. A recommendation nobody can execute is not a recommendation.
The handover
Backtests the team can rerun, drift controls, and documentation of every assumption that matters.
Selected work
Forecasting
Travel retail forecasting system
Multi-SKU demand forecasting pipeline for 30+ products across 100+ duty-free locations with automated monthly updates.
Forecast error: 38% → 24%
Pricing Strategy
Vitamins product pricing optimization
Data-driven pricing strategy for new vitamin product launch using Bayesian demand modeling and mixed-integer programming to optimize portfolio-wide pricing.
Revenue ↑18%
Pricing Strategy
B2B SaaS pricing optimization
Willingness-to-pay analysis and price sensitivity modeling for a B2B SaaS product, resulting in a new tiered pricing structure.
Revenue per customer ↑31%
Depth
Jul 25, 2026
Your data lake is not the prerequisite
Why data science and AI platform projects fail on definitions, not on infrastructure, and how to let a real decision pull the data model into existence instead of waiting for the lake.
May 8, 2026
Modeling Attributes, Not Products: Feature Engineering for Luxury Demand
What it actually takes to build a production forecasting system for luxury demand at SKU level: decomposing products into shared attributes, designing hierarchical aggregate features without leakage, handling intermittency with Tweedie and hurdle objectives, and forecasting in a way finance can use.
May 1, 2026
Forecasting Luxury Demand at SKU Level
A field perspective on hierarchical forecasting in luxury retail — where ultra-premium SKUs sell fewer than 100 units a year and finance cannot ignore them.
Apr 18, 2026
Designing a Markdown Optimization Engine with States
Markdown planning explodes when weeks, events, discount levels, and phases are modeled as separate dimensions. Collapsing each week into a single state turns it into a clean optimization problem.
How it starts
Diagnostic first, build second.
01 · Decision System Diagnostic
Three weeks, fixed scope. We establish which decision the forecast serves, whether the data can support it, and what the operating cadence needs to be. You get an executive readout and a roadmap.
02 · Design and build
Models, backtests, the decision layer on top, and the pipeline that runs it. Scoped from the roadmap, priced on the engagement.
03 · Operate and hand over
Monitoring, drift controls, documentation, and the training to run it without me.