Casino Enterprise Management

Authors’ Note: All numbers have been presented as percentages or indexes to obfuscate the raw data. Looking at actual figures may show different patterns.

In the previous articles in this series, we focused on the combined high-level results of the Penny Alley mini casino at Silverton Casino and the product-based marketing initiative that followed, talked about the opportunities that secondary games bring to the casino and most recently, the behavior-based clustering of Penny Alley customers.

This month, we’ll take a look at how behavioral clustering can uncover the hidden behavior in an otherwise similar group of customers, and how these similarities in behavior lend themselves to direct, product-focused marketing initiatives.

Last month, we used k-means clustering to segment Penny Alley players based on their product preferences. Using this method, we came up with six segments and determined their basic demographics, as described in Chart 1.

Chart 1

Product Choice Observations

Now let’s take a look at the product choices that created these clusters in the first place. To keep the data masked, we have used the letters A – H to represent different types of slot themes. These clusters are like market baskets (see “Turning Market Basket Analysis into Action” in the May 2009 issue of CEM). The heat maps in Graphics 1–6 show how the clusters of customer behavior are spatially distributed across the floor. [Note: The data and patterns represented in these graphics represent a distribution of time and volume for these select groups of players only, and may not be indicative of the overall revenue or performance trends in the area.]

Graphic 1: Cluster 0

Mal overwhelmingly prefers D themes, but she shows some secondary play on the A, B and C themes in the area as well.

Graphic 2: Cluster 1

Sasha plays A, B and C themes equally. She spends a disproportionate amount of money vs. time on E and F themes.

Graphic 3: Cluster 2

Lee spends more than half of his time playing A themes. He also plays the surrounding B and C themes and, like Sasha, spends a disproportionate amount of money to time on the E and F themes.

Graphic 4: Cluster 3

Jada prefers A themes but plays nearly everything in the area. She spends a wildly disproportionate amount of money to time on F themes and a moderately disproportionate ratio on all other themes, except G and H.

Graphic 5: Cluster 4

Ralph prefers B themes and spends disproportionate amounts of money vs. time on these themes as well. He shows a strong secondary preference for A and C themes.

Graphic 6: Cluster 5

Andrew prefers A and D themes equally but spends more money than time on D themes only.

Clustering to Drive Marketing Strategy

The Penny Alley area was initially marketed based on simple customer interest—any player who had either played in the area that was to become Penny Alley, or had played a product that was moving into Penny Alley, was invited. The entire group was treated the same, with the same mail piece, same message, same promotion, etc. But looking at the six clusters above, we can see they each have very different behavior patterns within Penny Alley and that each cluster displays different demographics. Using this information, we can now form a unique strategy to better target each of the clusters.

Cluster 0         

Since young Mal only plays D themes and the other clusters show very little crossover to D themes, chances are that a volume-based promotion created for the area left her out of the action most of the time. Should we create a specific promotion specifically for her on the machines she likes to play? Since the number of available D themes is limited in Penny Alley, should we consider including D themes in other areas of the casino to make the promotion more exciting and easier to participate in for Mal? Also, since we know that this group includes some of the younger casino patrons, should we adjust the promotion times and days of week to better accommodate their habits?

Cluster 1

What about Sasha, who has very even play patterns among A, B and C themes, but wagers more money on E and F themes during the limited time she plays them? Should we create an additional bonus layer for her on those E and F themes? And what is it about those games that make her increase her wagers? Since Sasha’s group shows even play patterns, the original promotion design should still work well here. One thing to address, however, is the moderate number of international members in the group. These members were not included in the original mail piece, but now that they are identified, perhaps we should share the details of the promotion with them during hotel registration or as part of their welcome packet.

Cluster 2

Then we have Lee, a bit older than Sasha, who really loves those A themes and who also shows the same behavior as Sasha when playing E and F themes. Why are the A themes such a favorite for Lee when Sasha doesn’t seem to care between A, B or C? And why do they both increase their wagers on E and F themes? Again, the original promotion design should still work well here, but with the group tending to be a little older, we should definitely make sure that the promotion accommodates their favorite time of day to play.

Cluster 3

Jada also prefers A themes but not as much as Lee does. She plays games from nearly all theme groups, including D themes, which we haven’t seen yet. Her spend-to-time ratio is the highest of all the clusters, especially on F themes. This group was markedly mixed in age, gender and proximity. With an increased spend compared to the other clusters and willingness to play all games in the area, a promotion with an increased frequency or reward amount might be a better fit for this group.

Cluster 4

Ralph likes the B themes and wagers more on these games than on any others. He’ll also play the A and C themes in the area. Should we add a layer to the promotion that rewards him more frequently on B themes, where he seems to have the most fun?

Cluster 5

Andrew is an out-of-towner who spends his time on A and D themes equally but who spends a disproportionate amount of money on the D themes. He also plays very few machines. Should we design a promotion for him that encourages him to explore more games? Or perhaps include him in the broader promotion but also make him eligible for Mal’s promotion when he plays D themes?

Cross-Sell Opportunities

Now that we know what our clusters of customers like to play, we can dig into each cluster and see if individual customers are getting the “full experience” that their cluster seems to enjoy. In addition, we can identify new customers’ likely clusters by their demographic and first visit behavior, and look to cross-sell them products enjoyed by that cluster.

There is data that indicates there is value in this approach. In Figure A, we map the number of visits per month by a customer against the number of theme categories purchased by that customer. When customers have very low frequency, they tend to sample more games as they try to identify which games they like. What’s interesting is that this “settling in” bottoms out at around nine visits per month—the truly loyal customers actually purchase more variety of products. Thus, if we can encourage customers to try more product types, we should be able to increase customer loyalty. Banks do this all the time, trying to get their checking customers to open savings accounts and purchase CDs in an effort to build brand loyalty.

In Figure B we see a linear relationship between time spent per visit and customer frequency, indicating that the customers increase their length of visit as they increase the number of products. Thus, the benefit to increased product purchases is to increase loyalty via the number of visits per month and to increase the actual time spent during each visit.

Now let’s do some examples with our clusters:

1. Existing Customer – An “Andrew” plays only A themes. Rather than trying to cross-sell Andrew on a wide array of different products, we can home in on his cluster’s preference of A and D themes. Thus our marketing cross-sell decision is to encourage Andrew to try D themes in addition to the A games he already enjoys.

2. New Customer – A middle-aged male starts playing at our property and chooses C-themed games on his first trip. He fits the demographic and the secondary theme choice of a “Ralph.” We can then predict that he is a likely member of this cluster and encourage him to play B-themed games, which is the strongest theme choice for Ralph customers and is likely a theme he will enjoy playing in addition to his C-themed games.

The Full Circle

So now we’ve really come full circle, starting with an analysis that drove a mini casino strategy, a combined slot and marketing initiative to promote the area using secondary games, analyzing those results and clustering the patrons of the area, and finally putting those clusters to work to further refine our product marketing strategies. As you can see, the options really are endless—and with the introduction of secondary bonus games by some gaming systems, creating rich multi-layered targeted promotions that take advantage of the information continues to get easier.

Andrew Cardno has more than 16 years of experience in business analytics, ranging from modeling health care drive times to casino gaming floor analytics. He often presents on the future of analytics across the world and has spent the last seven years living in the United States and working with corporations around the world. He can be reached at andrewcardno[at]yahoo.com.

Dr. Ralph Thomas is Vice President of Database Marketing for Seminole Hard Rock Gaming. During his years in the casino industry, Thomas has focused on maximizing profitability by applying statistical analysis to the company database. Previously, Thomas spent 15 years in academia, as both a student and a lecturer of mathematics. He can be reached at ralph.thomas[at]stofgaming.com.

Jada Evans is the founder of MindSight Analytics and works with gaming companies to provide analytics and help develop strategies for a variety of topics including game performance, marketing, food & beverage, hotel & labor. She can be reached at jevans[at]mindsightanalytics.com.