Cluster Analysis Reveals Hidden Insights

Taking customer segmentation to the next level

How many main segments of like-minded customers do you have in your casino database? Customers who share common attitudes, beliefs and behaviors?

If you have a tiered players club, you’re already segmenting your customers by how much they spend. In your direct mail program, you probably segment by recency, frequency and spend, at the very least. Some casinos have more than a dozen ways to segment their customers based on their tracked behavior in the casino.

The type of segmentation that casinos use for a tiered club or mail program is just one tool that casinos can use to understand their customers. Another type of segmentation can help you understand more about your customers so that your club benefits, advertising, promotions and customer service can all be custom-tailored to each segment in your casino.

This type of segmentation can be achieved with quantitative research of your customer base. An online survey can be developed that asks questions about several different areas, with a key area being what motivates the person to gamble. Data from the survey is downloaded and manipulated by advanced research software, such as SPSS or SAS. This software identifies and defines the characteristics of the main groups in your database in a type of analysis known as cluster analysis.

Cluster analysis has been around since the early 1930s and continues to be used because it makes it easier to understand large groups, such as a casino database. There are many casinos and companies using cluster analysis to identify clusters such as Escapists (gamblers who primarily visit casinos to escape from their everyday lives), or True Blues (gamblers who are highly satisfied and willing to pay a premium for their casino of choice), or whatever segments their database contains.

When SPSS or SAS software defines the number of segments in your database and the characteristics of each, it’s based on your unique survey data. There is no one set of standard segments and associated behavior that applies to every casino, so when you receive your analysis you can look at the characteristics that broadly describe the type of person in that group. Some casinos may choose to segment based primarily on behavioral characteristics, some may focus on attitudinal characteristics, and yet others may segment based on a combination of these characteristics.

There is a 2010 national study of U.S. casino gamblers that identified seven segments. The study used a sample from the Market Metrix Hospitality Index (MMHI). The MMHI is a national indicator of customer satisfaction, emotions, loyalty, and price sensitivity regarding hospitality industry facilities and services available to consumers. Study participants were people who had stayed at least one night in a casino hotel.

The final study sample of 4,894 responses was put through numerous iterations of cluster analysis in order to separate the data into logical segments that were sufficiently differentiated from one another. The eventual output produced seven distinct segments – three are summarized below, although the actual characteristics of each segment are more detailed. These segments were also defined using High, Medium or Low Visit Frequency as a main characteristic.

Segment Descriptions

High Frequency Visitors:

Elder Elites – This group represents about eleven percent of the consumer market and predominantly includes members of elite loyalty programs. These persons take advantage of these programs with a high number of trips. They are mostly male and have a relatively high income, but are considerably older than average. The Elder Elites are very satisfied with casino service, but not very optimistic about their perceived odds of winning, especially considering their loyalty and frequent visits.

Medium Frequency Visitors:

Unmoved Members – The Unmoved Members comprise an entire third of all casino customers. They are loyalty club members who take a few trips per year, but do not stand out very much with their behavior in contrast to other segments. Their income and spending are lower than other loyalty club members. They are more pessimistic about their odds of winning and less satisfied with casino service.

Low Frequency Visitors:

Happy-Go-Lucky – Accounting for nine percent of all casino guests, these customers tend to be younger, with modest incomes, who thoroughly enjoy their experience. They perceive their odds of winning as good, and express a willingness to recommend the brand they selected. Although they are not loyalty program members, they are quite willing to return again.

A significant finding of this cluster analysis is that for potentially very lucrative segments, loyalty and the willingness to pay a price premium are not always highly correlated with participation in loyalty programs or high visit frequencies. The Happy-Go-Lucky segment serves as the best evidence of this observation, since its members reported some of the highest satisfaction scores and a willingness to return and recommend their selected brand, without having already built a connection to the property or brand through frequent visits or the perks of a loyalty program.

ARTWORK:

Shown above is a depiction of how software such as SPSS can take all the variables in a survey – age, gender, spending, visit motivators, satisfaction factors, competitor visits, or whatever questions you want to add to your survey— and sift through them until it finds commonalities. The number of segments is based on the data at each casino. Each colored circle is a characteristic of that segment.

Deb Hilgeman 23 Articles