HP's OR stack added $117M to HPDirect.com's online revenue

HP entered online consumer sales with HPDirect.com, but rapid growth made marketing budgets, customer targeting and warehouse planning hard to run on judgment.

Operations research first identified and quantified the key drivers of online traffic to improve market planning and budget allocation.

Bayesian and Markov chain models then predicted which customers would buy which products, when, and through which marketing channel; hybrid forecasting set warehouse inventory and ensured timely fulfillment.

Integration of these solutions into HP's marketing planning and warehouse operations since 2009 generated an additional $117 million in revenue for HPDirect.com.

Predict who buys, then allocate spend to it

Quantifying traffic drivers stopped guesswork budget allocation.

Customer-propensity models targeted spend where purchase was likely.

Order forecasts aligned inventory with promised fulfillment.

A measured $117M revenue gain proved the stack's worth.

E-commerce growth hides waste: model traffic drivers, purchase propensity and orders as one chain so marketing spend and inventory follow predicted demand.

HP was named a 2012 Franz Edelman Award finalist for the HPDirect.com work; the Interfaces paper documents the additional $117 million in revenue generated since 2009.