Retail Price Optimization at InterContinental Hotels Group (IHG)
The Problem
Since their widespread adoption by the hotel industry in the early 1990s, revenue management systems have evolved to become a key capability to optimize occupancy and revenue. A fundamental assumption of these systems has been that demand by segment—particularly rate segment—is independent. In effect, industry leaders assumed that demand was independent of the price offered. Under this assumption, the benefits of revenue management derived from optimizing the allocation of inventory when demand exceeded capacity. Prices were predetermined via manual processes using little or no analytics. Most recognized the fault in the assumption of independent demand, but revenue management worked well enough under normal economic conditions.
The hospitality recession of 2002–2003, when the industry suffered occupancy declines of 10–15%, coupled with the growth and transparency of Internet distribution channels, motivated a reexamination of revenue management methods. IHG resolved to enhance its revenue management capabilities by relaxing the assumption of independent demand by segment and directly modeling demand as a function of price. In 2005, IHG initiated a series of projects that led to the development of a price optimization capability.
The Analytics Solution
IHG’s Retail Price Optimization is a significant, large-scale enterprise implementation of price optimization within the hospitality industry. This capability was designed to accommodate a globally distributed hotel organization with more than 4,000 users distributed with varied knowledge levels.
The key analytics components include a market response module and a price optimization module. The market response module computes price response models for each hotel. IHG developed various segmentation schemes and facilitated a hierarchical implementation allowing segments with insufficient data to be estimated using aggregate data. The core price optimization module is innovative and new to the industry. Modeling demand as a function of price and competitor rates requires that the objective function is nonlinear. Additional decision variables accounting for length of stay significantly increased the complexity of the model. A special optimization method was implemented and rates are generated in an automatic fashion based on hotel capacity, forecasted demand, current bookings, and competitor prices.
The Value
Retail Price Optimization has been rolled out to more than 2,000 hotels globally, and benefits measurement studies point to a 2.7% increase in revenue. This number was verified and acknowledged in the IHG 2009 Annual Report.
