1. BEST PRACTICE
Diamonds
in the
Data
Mine by Gary Loveman
Harrah's Entertainment has outplayed its
competition and won impressive gains, despite
being dealt a weak hand by the economy
The secret? Mining the company's rich database
to develop compelling customer incentives.
I
T'S A FRIDAY NIGHT on the Las Vegas she's happy with the service she's re-
Strip, and all of the neighbors are ceiving. Although the customer doesn't
making spectacles of themselves. The fit the stereotypical profile of a Las Vegas
$750 million Mirage boasts a Vesuvian high roller, Harrah's makes sure she
volcano that erupts loudly every 15 min- feels special. Because the casino delivers
utes. Next door, at Treasure Island, a the recognition and service she has
faux British frigate battles a pirate ship come to expect, she'll return to Harrah's
at regular intervals. Further down the again and again.
Strip, the Bellagio sports a lake fes- Harrah's Entertainment has the most
tooned with sparkling, dancing foun- devoted clientele in the casino indus-
tains that beckon to passing tourists. try-a business notorious for fickle cus-
Meanwhile, the customer pulling Into tomers. That loyalty to Harrah's, which
Harrah's Las Vegas is dazzled more by now operates 26 casinos in 13 states, has
the service than the building. A smiling enabled the gaming company to record
valet greets her by name. Instead of T6 straight quarters of same-store rev-
having to wade through a crowded enue growth. In 2002, Harrah's posted
lobby to reach the casino, she steps more than $4 billion in revenue and
quickly into the gaming room and sits $235 million in net income.
down at a slot machine. The card reader We've increased customer loyalty,
on the machine pages her host, who ap- even in the current challenging econ-
proaches every so often to ensure that omy, in two ways. First, we use database
MAY 2003 109
2. BEST P R A C T I C E • Diamonds in the Data Mine
marketing and decision-science-based
analytical tools to widen the gap be-
tween us and casino operators who base
their customer incentives more on intu-
ition than evidence. Second, we deliver
the great service that consumers de-
mand. In short, we've come out on top
in the casino wars by mining our cus-
tomer data deeply, running marketing
experiments, and using the results to de-
velop and implementfinelytuned mar-
keting and service-delivery strategies
that keep our customers coming back.
A Dicey Business
By the time I had left Harvard Business
School to Join the corporation as chief
operating officer in 1998, Harrah's had
become thefirstnationwide casino busi-
ness, thanks to a geographic diversifi-
cation plan promulgated by Phil Satre,
the company's chairman and then-chief
executive officer. Satre led the wave of
gaming growth in the 1990s, expanding
Harrah's from four casinos in two states
to 26 casinos across 13 states currently.
Satre's vision differed markedly from
the strategy pursued by other big casino
operators, whose "If you build it, they
will come" philosophy focused on at-
tracting customers to a fantasyland Las
Vegas. Companies such as Mandalay
Resort Group and MGM-Mirage in-
vested heavily in constructing costly
must-see casinos offering a wide range ness outside Nevada and Atlantic City, depend heavily on its stores, restau-
of amenities-fabulous spas, high-end seeing geographic diversification as an rants, bars, or shows; it drew the lion's
shopping malls, dazzling shows-de- opportunity to introduce the Harrah's share of its revenues-87.2% in 2001-
signed to appeal to a broader audience brand to new customers and to insulate from its casinos. He also suspected that
than simply gamblers. Their hope was the company from regional economic cultivating lasting relationships with
that such facilities would attract an ever- vagaries. Once Harrah's posted high re- the company's core customers-slot
growing number of new customers. This turns in these emerging markets, its players - would lead to greater and more
strategy ultimately transformed both competitors also began to expand into sustainable profit growth. So he opted
the Las Vegas skyline and tourist spend- them. But Satre believed that compet- to invest in development of the intel-
ing patterns: The Las Vegas Convention ing largely on the basis of hillion-doUar lectual and technological capabilities
and Visitors Authority reported that, in facilities in the face of new competition needed to assemble and analyze data
2001, revenues from dining, entertain- was not the most prudent use of capital about those customers. The goal was to
ment, shopping, and other activities because the returns on such buildings provide good service to them and thus
outpaced the city's casino revenues by often weaken when the novelty wanes. encourage their loyalty to the company's
a three-to-one margin. Fortunately, Satre had two important brand. When Satre hired me as COO, he
arrows in his quiver. First, he knew that, said he wanted to change Harrah's from
Satre, by contrast, focused on ex-
unlike its competitors, Harrah's didn't an operations-driven company that
panding the corporation's gaming busi-
viewed each casino as a stand-alone
business into a marketing-driven com-
Former Harvard Business School professor Gary Loveman is the CEO of Harrah's
pany that built customer loyalty to all
Entertainment. He is the author or coauthor of five HBR articles, Including "Putting
Harrah's properties.
the Service-Profit Chain to Work"(March-April 1994)-
lip HARVARD BUSINESS REVIEW
3. Diamonds in the Data Mine • BEST PRACTICE
One tactic the company had already tion into our enterprise data warehouse, develop just this kind of close relation-
decided to use to enhance customer which contained not only millions of ship with the people who visit Harrah's
loyalty was called Total Gold, a player- transactional data points about custom- casinos.
card program that was modeled after ers (such as names, addresses, ages, gen- Before we could persuade customers
the airline industry's frequent-fiier pro- ders) but also details about their gam- to come back time after time, however,
grams. Launched in 1997, Total Gold was bling spending and preferences. The we needed to take a hard look at them
designed to provide regular customers database was a very rich repository of and understand how much value each
with incentives to visit Harrah's prop- customer information. of them brought to us. We discovered
erties throughout the country. Custom- that 26% of the gamblers who visited
ers inserted their Total Gold cards into Slicing the Dicing Harrah's generated 82% of our revenues.
slot machines and earned credits as When we started digging into the data- We were surprised to find out who our
they played. They were rewarded with base, one statistic stood out: Our Total best customers really were. They em-
the standard fare that all casinos offer- Gold cardholders told us in surveys and phatically were not the gold cuff-linked,
free hotel rooms, dinners, show tickets, focus groups that they were spending limousine-riding high rollers we and oiu-
gift certificates. But there were three only 36% of their annual gaming bud- competitors had fawned over for many
problems with the program. First, noth- gets at Harrah's. This presented an years. Instead, they turned out to be for-
ing differentiated this program from opportunity. There was a promise of mer teachers, doctors, bankers, and ma-
our competitors' efforts. Our customers tremendous upside if we could induce chinists - middle-aged and senior adults
simply took their free rooms and din- customers to spend more of their gam- with discretionary time and income
ners and drifted across the street to do ing money at Harrah's and if we could who enjoyed playing slot machines.
their gambling. Second, our customers communicate effectively with them. I We also learned that thest? customers
earned different rewards at different suggested to Satre that we might be able typically did not stay in a hotel but vis-
properties; there was no uniformity in to divert more of our customers' annual
ited a casino on the way home from
the program. Third, and most impor- gaming budgets to Harrah's if we bor-
work or on a weekend night out. At the
tant, our customers were not given any rowed a page from the playbooks of
same time, we found that our target
incentives to consolidate their gaming other businesses whose case studies I'd
customers often responded better to an
with Harrah's. long taught. Basically, we needed to do
offer of $60 of casino chips thian to a free
what the Starbucks and Nordstroms of
While Total Gold wasn't much good
the world had done - change the way room, two steak meals, and $30 worth
for keeping customers loyal to Harrah's,
consumers made decisions about our of chips because they enjoyed the an-
it was quietly digging our future dia- ticipation and excitement of gambling
merchandise.
mond mine. By tracking millions of in- itself. And we were able to develop
dividual transactions, the information- To do that, we clearly needed to slice quantitative models that allowed us to
technology systems that underlie the and dice the data finely enough that predict, based on an individual's play,
program had assembled a vast amount we could develop effective marketing his or her "customer worth"-the theo-
of data on customer preferences. At the programs. Common practice calls for retical amount we could expect the cus-
core of the Total Gold rewards program defining marketing strategies apart tomer to spend not just during one eve-
(and its successor. Total Rewards, which from database strategies-that is, the ning but over the long term,
company comes up with a grand mar- Suddenly, we saw how we could dif-
keting scheme and then tries to adjust ferentiate our brand. Understanding the
Instead of focusing on how the database to its strategies. Unlike lifetime value of our customers would
much people spent in our many companies, we decided to let the be critical to our marketing strategy. In-
data suggest the specific marketing stead of focusing on how much people
casinos during a single visit, ideas to us.
spent in our casinos during a single visit,
it became clear that we The information we found in our it became clear that we needed to focus
database indicated that a loyalty strat- on their potential worth over time. For
needed to focus on their egy based on same-store sales growth instance, we could see that customers
would work. Same-store sales is a classic who said they were very happy with the
potential worth over time. measurement of a simple retail-loyalty Harrah's experience increased their
strategy: The goal is to get a customer spending on gambling at Harrah's by
I'll describe below) was a 300-gigabyte to visit your store regularly, just as she 24% per year; customers who said they
transactional database that recorded might routinely visit her hairdresser and were disappointed with Harrah's de-
customer activity at various points of mechanic. The hairdresser and the me- creased their spending by 10% per year.
sale-slot machines, restaurants, and chanic envelop the client in reasons to The best way to engage in this kind of
other retail areas in our properties. be loyal, primarily by developing a data-driven marketing is to gather more
Database managers fed that informa- friendly relationship. We decided to
and more specific information about
MAY 2003 111
4. BEST P R A C T I C E • Diamonds in the Data Mine
customer preferences, run experiments in line to park their cars, or eat in restau- we had signs on top of them proclaim-
and analyses on the new data, and de- rants, or check in at the front desk. So ing the odds? Our transactional data-
termine ways of appealing to players' we decided to make a point of routing base told us exactly what the patterns of
interests. We realized that the infor- our customers into three different lines. play were in our casinos. We discovered
mation in our database, coupled with People who weren't card-carrying Har- that at any given time, it was possible to
decision-science tools that enabled us rah's members and Gold customers know which specific customers were
to predict individual customers' theo- stood in lines at the reception desk or playing at particular slots in Harrah's
retical value to us, would allow us to the restaurant. Platinum customers Las Vegas and what it was about that
create marketing interventions that would stand in still shorter lines, and specific machine that appealed to them.
profitably addressed players' unique Diamond cardholders would rarely ever This knowledge allowed us to configure
preferences. The more we appealed to have to stand in line. This created a vis- the casino floor with a mix of slot ma-
these preferences, the more money the ible differentiation in customer service. chines that benefited both our custom-
customers would spend with us. It was essential for our customers to ers and our company.
So we decided to act on a radical idea: see the perks that others were getting.
We would reward customers for spend- Once we divided the lines this way, Hitting the Customer Service
ing in ways that added to their value. we watched as our customers did what Jackpot
Most consumer businesses insist that they could to earn the higher-tiered Deep data mining and decision-science
they can't treat one customer differently cards. Every experience in our casino marketing would be worth little in driv-
than they treat another, even though was redesigned to drive customers to ing same-store sales growth were it not
some customers are obviously worth want to earn a higher-level card. As it for another simultaneously applied and
tums out, marketing that appeals to cus- extremely critical ingredient-an ab-
tomer aspiration works wonderfully. solute focus on customer satisfaction.
All employees are told daily: We also set up a series of triggers in the When I came to the casino business,
If your service can persuade database and analyzed the customers' there was an insufficient focus on cus-
responses to those triggers. If, for exam- tomer service. We decided that great ser-
one customer to make one ple, we discovered that a customer who vice would allow us to build-just as
spends $1,000 per month with us hadn't Home Depot, Four Seasons, and other
more visit a year with us, visited us in three months, a letter or great brands do in their fields-the ca-
telephone call would invite him back. If pacity to brand ourselves as the only na-
you've had a good shift.
we learned that he lost money during tionwide consumer gaming business.
If you can persuade three, his last visit, we invited him back for a Customer service is something most
special event. Our telemarketers were organizations say they focus on. But, in
you've had a great shift. trained to listen for responses to specific fact, they often fail to institute systems
offers-a certain percentage of our cus- to use customer service to reinforce loy-
tomers responded positively to offers of alty with carrots and sticks. Our data
much more than others. To us, that ap-
a steak dinner; others would respond to told us that our customers want friendly
proach was fundamentally wrong, but it and helpful attention in addition to fast
offers of two free nights in the hotel.
didn't mean that we had to focus on the service. We decided to linl^ employee re-
relatively small number of high rollers. Once entered into our database, these wards to customer satisfaction. Accord-
Rather, we made a point of treating our responses provided fodder for more slic- ingly, we chose to measure all employee
millions of regular customers differently ing, dicing, and experimentation. It's im- performance on the matrices of speed
depending on their value to us. portant to note that our database strat- and friendliness. The better the experi-
It turned out that our customers-I egy hinged on our ability to combine ence the guest had, the more money em-
would venture to say all customers- data from all of our properties, so cus- ployees stood to make. To this end, all
actually enjoy aspiring to higher levels tomers could use their reward cards in Harrah's employees take part in a certi-
of achievement and reward. It's simply multiple locations. Combining transac- fication program that trains them to
human nature. Understanding this, we tional data from all our sites was so im- deliver excellent service. From house-
split our customers into three tiers: portant that we developed and ulti- keepers to slot attendants, from valets
Gold, Platinum, and Diamond card- mately patented the technology to do it. to stewards, from receptionists to chefs,
holders, based on their annual theoret- We also decided to use our transac- all employees are told daily as they ar-
ical value. Platinum and Diamond card- tional data to"seH"our slot machines. In rive at work: If your service can per-
holders receive greater levels of service, the past, we had no way of knowing why suade one customer to make one more
which adds an aspirational element to customers chose to play at certain ma- visit a year with us, you've had a good
the program. For example, our database chines. Was it because of the way the shift. If you can persuade three, you've
told us that our best customers wanted machines looked? Or because other ma- had a great shift.
service quickly - they didn't want to wait chines were occupied? Or was it because
112 HARVARD BUSINESS REVIEW
5. We implemented a bonus plan to re- Harrah's Laughlin earned the highest
ward hourly workers with extra cash for customer service scores in the company.
achieving improved customer satisfac- Our experience with customer service
tion scores, which we culled from very has shown us that meeting budget at
detailed customer surveys. If a prop- the expense of service is a very bad idea.
erty's overall rating rose 3% or more, If you're not mal<ing your numbers, you
each employee could earn $75 to $200. don't cut back on staff. In fact, just the
What has made the bonus program reverse: The better the experience a
work is that the reward depends on guest has and the more attentive you
are to him, the more money you'll make.
Meeting budget at the For Harrah's, good customer service is
not a matter of an isolated incident or
expense of service is a two but of daily routine. When he goes
off duty, the Laughlin general manager
very bad idea. If you're tells employees to call him at home-
not making your numbers, anytime, day or night-whenever they
seefivepeople waiting in any line. To us,
you don't cut back on staff. this is living proof that our same-store
sales growth in tough markets has been
everyone's performance. If the valet's driven by sustained attention to great
scores were low but the steak house customer service.
receptionist's were high, the reception- Wednesday, May 7, 2003
ist would check in on the valet. Likewise, The Winning Hand
12:00 - 1:30 p.m. Eastern Time
if one property received low scores and Everything we do to market Harrah's is
another high ones, the general manager framed in terms of players' decisions to
of the lower-scoring property might visit visit, or not visit, one of our casinos. One THE SUCCESS ATTRIBUTED
his colleague to find out what he could measure of the effectiveness of our strat- to Six Sigma at companies such as
do to improve his property's scores. egy is that many of our competitors have GE Is widely known. But not every
organization has similar results.
It's important to note that we chose adopted similar programs after viewing
How can your company unlock the
to measure customer satisfaction scores our company's performance over the transformative power of Six Sigma?
independent of a property's financial past few years. But our competitors con- How can you embed Six Sigma in
performance. In 2002, one property had tinue to focus largely on facilities, while your corporate DNA? What is the
record-breaking financial results, but we keep combining improved facilities relationship between leadership
employees did not receive bonuses be- with breakthroughs in marketing and and Six Sigma performance?
cause their customer service scores were customer service. We maintain our com-
mediocre. Our employees are obsessed Join Six Sigma expert
petitive advantage by using our human Subir Chowdhury, CEO of the
with their property's customer satisfac- capital and technology systems to get to American Supplier Institute
tion scores for a good reason: In 2002, know our customers better. Consulting Group and author of
we paid $14.2 million in bonuses to non-
Harrah's will keep adding excitement the bestselling The Power of
management employees based on their Six Sigma, and leadership
and benefits to the Total Rewards pro-
properties'customer satisfaction scores. authority Noel Tichy of the
gram, widening its scope across gaming-
Since the program's inception, Harrah's University of Michigan Business
related activities. And we'll keep en-
has paid out more than $43 million in School for a 90-minute interactive
hancing the benefits that players get
bonuses. virtual seminar in your office or
from consolidating their gaming within conference room.
This score-driven customer satisfac- our brands.
tion measure has allowed properties - Let the neighbors lure tourists with For details, visit
even those in troubled markets - to con- knights on horseback, fiery volcanoes, vvww.asiusa.com
tinue to grow. Take, for example, our pirate ships, and mini-Manhattans. We'll orcaii 1-800-462-4500
casino in Laughlin, Nevada. Despite just keep refining what we're already
strong competition and a mere i% in- pretty good at: drilling into our data and
crease in the local market's gaming making sure our regular customers are
revenues in 2002, Harrah's Laughlin more than satisfied. ^
recorded a 14% gain in revenues. Why?
Because its customers were loyal, thanks
Reprint R0305H; HBR OnPoint 3647
to great service. In fact, the employees at To order, see page 131.
MAY 2003