Sunday, August 22, 2010

Harrah's Loyalty System

A Data Driven Implementation of The Loyalty Effect in a Retail Setting

Gary Loveman became a professor at Harvard Business School after earning a Ph.D. in Economics from from MIT. His research emphasized "that businesses need to identify their most profitable customers and get them to shop more often." After penning an influential article on the topic, he consulted for Harrah's Casino, later submitting an unsolicited letter of suggestions to the CEO... culminating in his being offered the vacant COO position and his succcesion to CEO five years later.

His timing was exceptional, as Harrah's had already invested $17 million on customer tracking and data storage by providing them with loyalty cards. They had also patented a tracking system that incorporated demographics data, gambling history, and customer preferences. Armed with this data and his knowledge of quantitative methods, he quickly identified the profitable customers, business drivers, and debunked the prevailing Vegas business strategy.

First he realized that Harrah's made more money from elderly slot machine players than the high-rolling whales traditionally courted by casinos. The data showed that every casino was competing for the same 500 whales, while the highly profitable slot machines were being ignored. This led to another insight, that traditional Vegas business models spent billions on the construction of ever more luxurious casinos to attract these whales, but they only produced profits in their first year... quickly fading into financial problems. Not only did Harrah's shun this undesirable business model, they profited from it by acquiring the insolvent properties at distressed prices... rather than building them from the ground up.

What Is a Gambler Worth?


After answering the high level question of "Which Customer Segment Is The Most Profitable?" Loveman asked "How Much Is Each Customer Worth?" But rather than produce results at an aggregate level, they sought to pinpoint the exact value and potential value of each individual customer... based on demographic data, gambling history, customer preferences, and sensitivity to Harrah's marketing.

For example, "females are more regular gamblers than males and older women are the most lucrative demographic of all."  Then they estimate income based on public data and betting patterns. Some people bet rarely and only on games with better odds, while others burn money. Harrah's seeks to reward the more profitable customers with comps. A simple idea yes, but historically it was the "flashiest high rollers and the noisiest comp-seekers" being rewarded... rather than the most profitable customers.

Data Driven Marketing


Harrah's also estimates a gambler's available spend, and then markets in real time to that customer to earn the highest possible percentage of that available spend. If customers stop visiting, "Harrah's continues to pitch specials to them, but it shortens the deadline. And when loyal customers--tired of losing--consider skipping to another casino, Harrah's makes a preemptive strike: Spend more and be rewarded with a steak dinner or free cash the following day." Similarly, "A customer may, say, come into Caesars Palace and spend $1,000 per hand on blackjack for 15 minutes and leave. 'That person's spend with us might not seem to justify any reinvestment in him, but we can see [his demographic data] guess that we're getting just a little bit of his gaming budget.' The customer may then get a marketing offer that's worth more than the revenue Harrah's collected from him in those 15 minutes.' High value customers are always identified as quickly as possible and given preferential services, meanwhile unprofitable customers are largely ignored. They also send out thousands of surveys every week, and use the responses to predict what customers want... whether it be free steak dinners or spa treatments.

Harrah's also excels at managing their capital investments to encourage gambling and overall profitability. "We don't fill our rooms with bodies, we fill them with gamblers." They accomplish this with the use of a $7 million dollar "hotel yield-management system, which gave reservationists the ability to decide what to charge each customer for a room... in such a way as to optimize profitability... The company also acknowledges the sophisticated management of new slot machines and has recently focused on valuing the impact of hotel room remodeling and restaurant selection.

Applications of Quantitative Methods


  • Identify the drivers of profitability in the industry
  • Compare different business models, and choose optimally
  • Calculate incremental customer profitability
  • Integrate demographics and calculate customer available spend
  • Calculate Lifetime Value (LTV) of each customer 
  • Optimize hotel utilization for lodging the maximum total gambling loss
  • Integrate survey data with customer profiles and then predict customer preferences
  • Maximize profit from capital investments
  • Evaluate the effect of marketing programs on customers and to optimize marketing spend

Performance Analysis