Measuring the player experience: what a squiggly line can tell you

Why people gamble has long been debated. That people gamble is undisputed: We’ve done so all through recorded history. One of my favorite personal toys is a single die (two or more die are called dice), carved from bone back around 300 B.C. Except for its yellowed appearance and less than perfect cubism, it could be put to use on modern craps tables and barely noticed-witness to how little gambling has changed in 2,300 years.

Each gambling transaction culminates in either a win or a loss; while the former is preferable, the latter provides meaning. Without real possibility of loss, winning cannot gratify because gambling is more about overcoming loss rather than material gain. Loss makes the business of recreational gambling possible; if players do not collectively lose more than is won, there is no profit.

Within the human psyche, joy of winning far outweighs the cost of losing. Humans are programmed to endure hardship, setback and disappointment in the pursuit of occasional accomplishment. In our work, our sports, our politics and our wars, occasional winning outcomes justify any required sacrifice. In fact, a series of costly losses sometimes enhances the importance of a win. Gambling is simply a microcosm-an entire cycle of life-lived out in the space of an evening or an hour.

The wagering experience is so compelling that, since shortly after the invention of gambling, governments have tried to limit, or eliminate, the enterprise. Many of religious, or moral, inclination work hard to ban the practice, not only for themselves, but for everyone, on the theory that society’s very fabric is damaged by pursuit of unearned accomplishment.

In fact, legal barriers to gambling have made the casino industry, as we know it, possible. Without such barriers, gambling machines would exist in every bar, every grocery and every restaurant. Tens of millions of machines, each earning only a few dollars a day, would populate the Earth. Only because of opposition to gambling are large numbers of games concentrated in a relatively few locations where they are better controlled and taxed.

So effective are these barriers, according to the American Gaming Association, that, as popular as gambling has seemingly become within the United States, just 1 in 4 adults even set foot in a gambling establishment last year. Just 20 percent of those visitors accounted for some 90 percent of total wagers.  In other words, nearly all wagering revenues come from barely 5 percent of the adult population. AGA surveys show that visitation, as a percentage of population, has held steady for at least the past 10 years.

Yet the U.S. adult population is growing at just 1 percent annually while casino game population is growing much more rapidly. As I’ve noted in previous columns, casino expansion has been so prolific over the past decade that game supply now exceeds player demand. Our industry, which is built upon rapid expansion, cannot continue to grow-unless we convince a larger percentage of people to engage in the gambling pastime.

Convincing people to gamble requires a lot of effort-or at least a lot of new effort-as compared to current marketing campaigns, which aim largely at getting existing players to return more often or pirating other casinos' customers.

Before the work of bringing in new players efficiently begins, let’s make sure the people that do visit us are largely satisfied with the experience. If we can engage a larger percentage of visitors, work done to attract new customers will be more easily amortized.

How can we convert more visitors into loyal customers? Just asking that question is a great start because, bluntly, we’ve not paid much attention to this opportunity and a variety of good possibilities exist.

Here’s one: A variety of surveys confirm that if a new player enjoys a winning experience on his or her first outing at a casino, that player is far more likely to become a loyal customer. Anecdotal evidence further shows a player who doesn't win with some frequency will eventually grow frustrated and stop playing. While that seems counter to my earlier comments that a streak of costly losses amplifies the value of a win, we all have limits.Those limits vary from person to person and to the mood of a person at a given time.

How can we figure out what these limits are? Let’s start by measuring player experience and response. Right now, our systems focus primarily on accounting, security and basic rewards. We simply don’t try to quantify player experiences, much less determine what sort of experiences players best respond to.

One way to measure the quality of a player’s experience at a gaming machine is through a method I call “Squiggles.” The dictionary defines squiggle as “a short twisty line.” In the technique I’m about to describe, a squiggle indicates an enjoyable session of gambling.

Figure 1 graphs an imaginary play session. The vertical axis shows the credit meter balance on a game, the horizontal axis represents games played. In Figure 1, a player inserts $20 (red arrow line) and wagers $1 per game. She plays 20 successive games and never wins. Her credit balance, which began at $20, drops by $1 at a time until, after 20 games, there’s no money left. The graphed line is straight and has a constant negative slope, meaning the line goes downward as successive games are played. There’s no twisty part of the line because the player always loses. The casino is happy with the result: $20 won. Our player isn’t so pleased and is less likely to be a long-term customer.

Figure 2 graphs a second imaginary play session. Again our player starts with $20. Again she wagers $1 per game, but this time, she wins $10 on each wager. Her credit balance grows by $9 per game, and certain her luck can’t continue, she cashes out (red down arrow) after 10 wagers with a $90 profit. She’s happy, but casino management certainly is not.

Players who only experience losses (Figure 1), eventually give up and never return. Those who experience constant winning outcomes (Figure 2), will return-but not for fun-they’d simply come in to collect money-a job of sorts.

 Imagine a slot machine paying 102 percent  with no volatility, meaning the player wins $1.02 on every $1 bet. If each game lasts 6 seconds, a player wins 12 cents per minute or $7.20 per hour. Game play is now a low-wage occupation. Without volatility-the chance to win, with a risk of loss-gambling is simply not interesting, even when payouts are more than 100 percent guaranteed.

Now consider Figure 3. Again our player begins with $20 of credit. Again she wagers $1 per game. Each of the first three wagers are losses, followed by a $5 win, six losses, a win of $2, eight  losses, a win of $5, six losses, a win of $2 and seven more losses. Notice the squiggles in this graph. They mark the times of change. Losing sequences turn to winning ones and vice versa.

Squiggles mark changes of fortune and are the reason players play. I am not saying the Figure 3 squiggles define a satisfying player experience. (For our purposes, let’s define a satisfying experience as one that would cause the player to gamble again in the future.) I am saying that play experiences without squiggles are unsatisfying.

Exploration is needed to determine how big, and how far apart, satisfying squiggles must be. I suspect it will vary significantly by player personality and circumstance. If we were to measure player experiences this way, we could use the information to improve game offers.

For example, suppose we collected initial play information on 1,000 first-time player experiences: people that signed up for club membership and then gambled at least $50. Next we bin these players into two groups: those that returned to gamble again at a later date and those who did not.

Finally, we average the play experience graphs of each group. I’m betting you’ll find answers when these are compared. I believe the return group will show steeper squiggles. I’m betting, too, that we can learn how to build better games and use bonuses to greater effect if we learn this tool.

Besides just the squiggles, we can learn a lot by watching for cash-outs or the insertion of additional money and what the preceding experience was. For example, if a player inserts $20 and cashes out after spending only $10 she could be bored-or maybe she just won a $100 jackpot. Graphs and squiggles would help us understand such details of behavior.

One big problem stands in the way: raw data. Today’s systems simply don’t collect and store play-by-play game information. That’s a shame because modern networks have plenty of bandwidth, and disk storage space is cheap.

Closing comments

As much as I’ve enjoyed writing these columns, this one is my last. My goal for 2009 is to write with a deeper analysis of player behavior and measurement. Columns just aren’t good forums for such goals. Instead I hope to put these ideas into a book.  Though the audience is limited, the topic is too important to leave unexplained.

Thanks for reading.