by Adam Tanner
14
Seeking the Goldilocks Balance
Fostering Loyalty
The best businesses give consumers a choice whether or not to share their data, and offer benefits in return. The trick for executives is to strike the right balance when offering incentives. Big data helps companies such as Caesars tweak their formulas to best influence customer behavior.
People will share information for discounts or rewards, a trend that has accelerated in the Internet era. For example, some car insurers offer better drivers lower prices if they allow monitoring of their driving habits. So motorists agree to install a telematics device in the car’s dashboard, which sends back information such as what time of day the person drives, how fast he brakes, and other characteristics. At a company like Progressive, about two-thirds of drivers in the program get a 10–15 percent discount. So far, the company has avoided charging bad drivers more because it has been encouraging people to sign up. But in the future, bad drivers will have to fork out perhaps a 10 percent premium.1
At Caesars, data from loyalty cards set the reward levels to motivate how people spend money. That’s why Dan Kostel received an offer for $1,000 in free chips and a free room and meals at Caesars Palace. It seemed he had hit the jackpot even before crossing through the casino’s always-open doors. It wasn’t, however, a case of good luck.
Caesars singled out Kostel because he looked like the kind of player who would spend a lot and because he appeared receptive to their offers. They looked at a dozen or more characteristics such as how far away he is from Vegas, his average bet, his favorite game, the things he charges to his room, and how often he comes. If he has not come in a while, he may get a particularly enticing promotion. Once he starts coming regularly, the offers may diminish. If the company calibrates its offers properly, both Caesars and the customer end up happy.
Kostel estimates that he lost a total of about $5,000 at the blackjack table during his first year as a Caesars regular. He knows that gamblers sometimes underestimate their losses and concedes the same may have happened to him. In any case, he says the casino gave him at least $5,000 in free rooms, food, drinks, and other perks, so overall he is satisfied—as is the casino. CEO Gary Loveman said Kostel likely helped the casino’s bottom line, even though he received many freebies. “It is very unusual that someone whose actual loss in gaming is $5,000 or perhaps a bit more, for that not to be profitable to us,” Loveman says. “That’s how it should work: we should be able to give you things that you care about—not have you littered with things you don’t care about—and have it work out profitably for us.”2
Learning More About Customers
Customers today may allow an electronic “little brother” in the car or carry around loyalty cards in exchange for discounts, but once upon a time, people held their personal information close. They warily eyed merchants who seemed too nosy and asked too many questions. Businesses got used to learning a little about customers and then supplementing their files with outside data.
Veteran marketer Barbara Eskin has noticed a shift in the public’s willingness to share information. She started her career in 1979 at Ladies’ Home Journal. Back then the magazine found subscribers by renting subscriber lists from rivals such as Good Housekeeping and Family Circle.3 “People were much more cautious about the information. It wasn’t as easily accessible and, moreover, people thought it should be their own personal information,” she says.
Over the years, Eskin saw growing sophistication in how much marketers could learn about customers. In 1983, the US Postal Service introduced ZIP+4, which gave marketers far more precise information than before, sometimes pinpointing a single large apartment or office building. Marketers started mapping out micro-neighborhoods with customer information that became easier to buy over time.
More recently, Eskin worked as director of the loyalty program at high-end chocolatier Godiva.4 Like Total Rewards at Caesars, Godiva’s program offers incentives to tease out customer information that could help drive more sales. Godiva offers a free piece of candy every month just for signing up, which initially meant volunteering basic details such as name and email. Members can collect their rewards, worth $2.25 to $2.50 retail (although it costs just a fraction of that to manufacture), at any US or Canadian Godiva store. More than five million people joined in the first few years after the program’s 2009 inception. Yet it did not work as widely as Eskin had hoped. Many did buy more chocolate, but 1.1 million signed up and then disappeared. The company needed more information to understand the people who were interested enough to sign up but who did not become good customers.
Eskin discussed Godiva’s mixed success with other marketers at a Direct Marketing Association training seminar.5 Godiva could ask for more information at sign-up, but Eskin did not want to deter people from joining by demanding too many details.6 Seminar leader Devyani Sadh, CEO of the marketing firm Data Square, suggested big data could help. Godiva could learn a lot more about an email address by buying supplementary personal information from data brokers. She paused for a moment and then described the type of companies Eskin should seek out: “reputable.” Everyone chuckled. The field has a mixed reputation, even as data appending has become increasingly commonplace.
Data appenders vary in size and stature, from massive data brokers such as Experian and Acxiom to mom-and-pop startups. Acxiom says such data make a big difference. Placing ads in the right context—for example, showing shoes on a fashion page—boosts sales 32 percent. Demographic information such as name and address can boost performance more than fivefold. Adding behavioral data such as what sites people are looking at in sequence to determine their likely buying habits (the focus of Chapter 13) lifts responses more than twelvefold. Adding all this information together rockets response rates fifty-five times more than untargeted media, Acxiom tells clients.7
Chicago-based Rapleaf is an example of a smaller data appender. For a penny per piece of information, the company adds age or gender to an email address, or other details such as location, income, marital status, presence of children, homeowner status, or a person’s areas of interest or past purchases. Some marketers send in lists of as few as a thousand emails, says Rapleaf CEO Phil Davis. Others send millions. When we last spoke in 2014, he said the company had handled more than three billion emails over the past quarter.8
Davis gave an example of how additional data might help a website offering daily deals. For a special on manicures and pedicures, just knowing age and gender could shape a different appeal to married men. “They could say, you know, ‘If you are like most men, you’re going to screw up and your wife is going to get mad at you. Next time don’t give her flowers, have this mani-pedi ready for it. Oh, by the way, don’t be a Neanderthal, your wife might like it if your nails looked good once in a while,’” he says.
Davis admires how Gary Loveman and his team at Caesars Entertainment use personal data to target offers, showing the company’s impact far beyond casinos. “Caesars is awesome,” he says. “It is one thing to single out one data field and one result, but they put it into an algorithm. I think frankly that analytics is a really cool thing. Take lots of data, put it into an algorithm, and come up with a better recommendation. So now the recommendation could be a specific club. Here’s the good club on Tuesday nights you should be going to.”
Caesars, whose casinos include the Harrah’s, Bally’s, and Horseshoe brands, extending as far as Britain, Egypt, and South Africa, have detailed insights on likely customer spending because they track not just betting but hotel stays, meals, and many other things for which the more than forty-five million loyalty members have spent money.9 Over many years, Caesars and rival casinos have figured out someone’s likely future spending by simple math: average bet, multiplied by bets per hour, multiplied by total time gambling, multiplied by the house’s mathematical advantage in the gambler’s preferred game. That equation means a slot player betting the same amount over the same amount of time as a good bl
ackjack player may be more valuable because the odds are worse on slot machines. At Caesars, about 85 percent of gambling revenue comes from Total Rewards members, making the loyalty program what Loveman calls “the big dog in the company.”10
By profiling their best customers, Caesars want to find more people like them. Who is the next Dan Kostel, ready to travel from Los Angeles to Las Vegas once a month to gamble thousands of dollars? Who are the New Jersey retirees who might visit Atlantic City dozens of weekends a year? For some companies, identifying likely customers is relatively easy. A manufacturer of cribs and diapers looks for new parents. Landscapers, alarm companies, and Internet and cable providers rent lists of people moving into new homes. Other goods and services, such as gambling and fine chocolate, are harder to target. Caesars executives long felt they possessed the most relevant information to boost their business by watching what customers do in their hotels and casinos. But the financial crisis of 2008 jolted their forecasting models. A financial crisis makes people think hard about how they spend their money. Some changed long-established habits.
“Let’s imagine Adam Tanner used to be part of a segment of people that behaved largely the same way—guys your age—buys the same sorts of things at the same rate. Since the crisis the divergence of behavior in your group has gone way up,” Loveman says. Once-similar customers were no longer acting in a herdlike manner. “So if you were selling pools you’d find a guy who bought a pool and you’d go to other neighbors and say, ‘Look, the Tanners have a pool, you’ve got to have a pool.’ What we observe recently is that it is much harder to predict what similar people will do after the crisis.”
Changing the Magic Balance
Caesars had another problem. Company executives felt they were, in effect, paying too much in perks and benefits for customer data. Caesars give back about $25 in perks such as free rooms, food, and chips for every $100 of profit. Of that amount, about $2.50 funds points that a customer can redeem for free play, food, entertainment, and hotel stays. That’s the advertised part that gamblers know they will be getting. On top of that, the company allocates an average of $22.50 per $100 profit for unpublished targeted offers and comps—such as the free chips offered to players like Kostel.11 Those offers are not democratically distributed. Valuable players may receive freebies worth $40 per $100 of company profit. Others might get just $5—based on what Caesars believe will generate the best return.
Caesars gave out a lot of free food and drinks in their loyalty club lounges, but fretted some people stayed too long and consumed too much given how much they spent over the course of the year. Behind their backs, staff dubbed such guests “grazers.” One homeless man spent much of his $1,400 monthly disability check at the Harrah’s in Kansas City to earn Diamond status so he could eat every meal there for free. He was eventually kicked out after, in effect, figuring out how to squeeze the maximum value of the loyalty program to his benefit.
Because of tougher times since 2008, fewer Caesars clients were working their way up the tiers of Total Rewards. Caesars hoped their members would aspire to move up from entry-level Gold to Platinum, Diamond, and then Seven Stars status—and spend more throughout the process. But the progression wasn’t happening the way it once did. The slow economy was not the only problem. At the start of Total Rewards, Caesars had unmatched technology and expertise. Over the years, rival casinos introduced their own popular programs. Some of Loveman’s math nerds defected to the competition.
Caesars had to rethink their approach, so they decided to change the parameters of Total Rewards. It was the type of adjustment airlines have embraced from time to time. Other programs are also constantly tinkering to find the “Goldilocks balance”: enough free chocolate, flights, meals, hotel rooms, and other perks to keep people signing up with their personal data and coming back for more—without hurting the bottom line.
Total Rewards chief Joshua Kanter worked throughout 2012 on the program’s most significant revamp in more than a decade. The new plan required more points to reach the next level but gave bonus points for big spending on any single day. By raising the lowest-level elite Diamond threshold from eleven thousand to fifteen thousand points, the company hoped to thin the ranks of grazers. Kanter calculated that 10 percent of Total Rewards members would fall to a lower status. Yet perhaps 25 percent fewer people would crowd the elite lounges, creating a more appealing atmosphere and saving the company money. He felt overrun in the Diamond lounges, the second-highest level in the Total Rewards program. Sometimes twenty-five people were lining up to get into the elite lounge—hardly VIP service.
***
On a fall afternoon in 2012, Kanter took some deep breaths ahead of the first in a series of conference calls in which he was to explain the changes to top management across the country. Although he kept his feelings to himself, he had mixed emotions. He had good reasons to tinker with a successful formula, but he knew any change risked inciting a New Coke debacle. “In one ear I have operational leadership saying . . . ‘I need your help to focus our property operations on the very best customers,’” he said. “In the other ear, I’ve got general managers from across the network saying, ‘I can’t believe you are doing this. You are going to completely piss off my best customers, who come every single day.’”
Kanter was still relatively new at Caesars, but he had been around long enough to know that the changes would bewilder some managers. “Any time there is change there are people who freak out because it’s too much change, because it’s too confusing and they don’t think we can execute on it,” he thought. “And then there are other people who will say, ‘It’s not big enough. Why are we going to go through all this trouble?’”
Kanter dialed into the conference call and started his presentation. Managers reacted cautiously at first. Gradually, one after the other aired concerns about the impact on the eight million active Total Rewards members—the core audience vital to the company’s future.12 Darrell Pilant, a vice president of operations and marketing in San Diego, worried about how the program’s best customers, its Seven Stars members, would react after learning they would need 50 percent more points—150,000 points from 100,000—to qualify for that top-level elite status for 2014. “I’m just thinking about the difficult conversation,” he said. “The 50 percent increase is going to be a slap in the face.”
John Koster, the regional president of Harrah’s in northern Nevada, later outlined his concerns in an email.13 “Our competition is not raising the play bar to maintain loyalty program status,” he said. “They possess newer products with excellent locations (Indian Gaming) and aforesaid aggressive reinvestment. Add to that the rotten economy . . . and this is certainly somewhat of a risky initiative. Loss of Tier Status is ‘loss of face.’ Customers don’t take kindly to that and tend to show their angst via lost trips and play, or complete defection.”
Kanter’s direct boss, Tariq Shaukat, the executive vice president and chief marketing officer, thought the internal unveiling did not go as well as he had hoped. He believed they could have done a better job of explaining why they were making the changes. Over the next week Kanter held another five conference calls, talking to a total of a few hundred people. Each time, he grew better at explaining the complicated changes. Calming fears of both Caesars staff and Total Rewards members was so vital that both Kanter and Shaukat—who, like Kanter, joined Caesars from McKinsey & Company—personally drafted the message to Total Rewards members.
Another problem threatened before the announcement went public. Some of the management team were sharing details of the upcoming changes with customers. Partial leaks appeared on blog sites. In early December the company sent around a stern warning: “We put ourselves, our players and this great program at risk if we do not roll the information out properly,” Brandi Ellis, senior vice president of VIP marketing, wrote in a memo.
The Carrot and the Donkey
Loveman was not involved in the day-to-day changes, but he watched from the sidel
ines. He knew moving the goalposts would make some people very unhappy and spark debate inside the company. The people who faced the greatest challenges were the casino staff who had to explain the changes to guests. “These are huge high-volume operating businesses, so when you mess with something like Total Rewards there’s going to be thousands of people coming to our staff in the property with an issue,” Loveman said. “If you get it wrong, you make their life miserable.”
People did complain. Rodney Holland, a regular video poker player at Harrah’s in Kansas City, reacted warily when he first heard about the changes. “They had better not move the carrot too far ahead of this donkey, because then this donkey may not come at all,” he huffed. Holland had been frequenting Harrah’s Kansas City since it opened in 1994. He did not spend huge amounts when he gambled—typically $189 on average. But he came at least once a week, sometimes four or five times in one week, enough to earn him 138,000 tier points by late 2012—well clear of the old 100,000 level that bestowed Seven Stars membership, but short of the new level of 150,000 points.
A few months into 2013, Loveman thought the Total Rewards revamp had gone fine, although some officials remained worried. Clients were changing their behavior, which was the goal in the first place. They gambled more on the days they came, because the new rules gave players bonus points for spending more on any given day. Yet since customers spent more when they came, they were coming less frequently. Average daily spend went up, but the number of visits went down. Overall, they spent about the same amount of money as before, although some key clients spent more.
“That is actually what we had designed the strategy to promote, and that is, in fact, what we are seeing,” Kanter said. Yet the decreased frequency of visits alarmed some company officials. “Now we’ve got people who are freaking out around the company, about, ‘Oh, my God, visitation has gone down, what am I going to do?’”