Gambling – the wagering of money or something of material value on an event with an uncertain outcome with the primary intent of winning additional money and/or material goods – has been with us since ancient times. Greek mythology tells the story of Poseidon, Zeus and Hades dividing the world between them in a dice game; Poseidon won the sea, Zeus the heavens and Hades the underworld. The land, I suppose, was left to the rest of us.
Gambling has, of course, changed dramatically since those ancient times. Today’s casino operators are faced with a gambler who is much more sophisticated than the ancient Roman soldier who tossed a coin in the air and called “Heads or Ships.”
To succeed in today’s highly competitive gaming landscape, casino operators must understand their customers like never before. Luckily for them, we have entered a brave new world of gambling and entertainment where customers knowingly – and oftentimes unknowingly – leave clues to their gambling behavior. Player cards allow casinos to get a deeply personal view of their patrons, a view that can reveal not only the patron's gambling habits but also their dining, spa and shopping spend throughout the casino property. Predictive analytics can analyze this data, quantify every dollar a customer spends on property and then predict the customer’s unique value to the casino, while also providing clues that will make marketing campaigns directed to this individual much more effective than scattershot direct mailing campaigns.
Intensive competition in the gaming industry is making it more and more important for casino operators to not only pinpoint the small percentage of players who make up a large percentage of their profits, but also to create a long and lasting relationship with these players. By learning about a customer’s unique wants, desires and needs, a casino property can market directly to those wants, desires and needs, thereby making their marketing efforts much more effective.
Double Down On Your Data was written to show casino management how to best cull through their in-house patron data to discover who their most profitable patrons are, and also to show them how to market to these individuals to create a long, lasting and highly profitable relationship.
Double Down On Your data will help casino executives understand all of the current tools available to casino operators; tools that can help them understand their patron's better; tools that can help them create better prediction models; tools that can help them market to their players more effectively; and, most importantly, tools that can help them raise their casino's ROI.
Gambling has been with us since ancient times and it will undoubtedly be with us forever; it is just too ingrained within the human psyche to ever go away. Double Down On Your Data concludes with a description of an ideal solution that gives casino managers a clear understanding of their patrons. A solution design
3. Predictive Analytics for the Gaming Industry
“In business, as in baseball, the question isn't whether or
not you'll jump into analytics. The question is when. Do you
want to ride the analytics horse to profitability...or follow it
with a shovel?”
-- Rob Neyer, ESPN
4. Predictive Analytics for the Gaming Industry
Predictive analytics refers to a variety of statistical
techniques that analyze current and historical facts to
make predictions about future events. Using such
techniques as predictive modeling, machine learning,
data mining and game theory, predictive analytics can
build models that exploit patterns found in historical
and transactional data to identify business risks and,
just as importantly, opportunities.
5. Predictive analytics can be broken down into three
different types of models:
Predictive: these analyze past performance to predict
the likelihood that an individual customer will exhibit a
specific behavior in the future.
Descriptive: these identify different relationships
between customers to group or segment them for
marketing or other purposes.
Decision: these predict outcomes of complex decisions,
relationships, products and/or processes.
6. Predictive analytics extracts information from data
sets and uses it to anticipate future trends and
behavior patterns based on statistics and data
mining (Ramakrishnan and Madure, 2008). The
most important element of predictive analytics is the
predictor, “a variable that can be measured for an
individual or other entity to foresee future behavior”
(Ramakrishnan and Madure, 2008). The real trick is
to find the predictive model best suited for the
outcome one is trying to study (Ramakrishnan and
Madure, 2008) and this is no easy feat.
7. Predictive analytics solutions include SAS's suite of
analytics products, IBM's SPSS, EMC's Greenplum
and Revolution's R open source product. Whichever
solution is used, predictive analytics can enhance
customer acquisition and retention, identify cross-
sell and up-sell opportunities, identify customer
lifetime value, spot fraud detection, determine the
life cycle of a slot machine and help direct and
improve marketing campaigns.
8. Data Mining: An In-House Goldmine
Data mining – the process whereby hidden patterns
within data sets are discovered – is a component of
predictive analytics that entails an analysis of data to
identify trends and patterns of relationships among
data sets (Ramakrishnan and Madure, 2008). To put
is simply, data mining helps transform raw data into
usable information.
9. Data Mining: An In-House Goldmine
By employing automated predictive analytics to sift
through a casino operator’s customer database,
data mining can discover hidden opportunities and
connections that might otherwise be missed. Many
casino operators have terabytes and terabytes of
data – everything from customer player card
information to information about a customer’s room
preference – and sifting through this information to
discover meaningful connections would be an
impossible task without data mining
10. Data Mining: An In-House Goldmine
Data mining and predictive analytics aim to identify
valid, novel, potentially useful and understandable
correlations and patterns in datasets (Chung & Gray,
1999) by combing through copious amounts of data
to sniff out patterns and relationships that are too
subtle or complex for humans to detect (Kreuze,
2001). Data must be gathered from disparate
sources and then seamlessly integrated into a data
warehouse that can then cleanse it and make it
ready for consumption.
11. Data Mining: An In-House Goldmine
Trends that surface from the data mining process
can help in monetization, as well as in future
advertising and marketing campaigns.
For casinos, data mining can cull through data from
such disparate sources and departments as sales
and marketing, thereby allowing users to measure
patron behavior on more than a hundred different
attributes, which is a far cry from the three or four
different attributes that statistical modeling used to
offer.
12. Applications for Predictive Analytics
Cross-sell/Up-sell – 47%
Campaign Management – 46%
Customer Acquisition – 41%
Budgeting and Forecasting – 41%
Attrition/churn/retention – 40%
Fraud Detection – 32%
Promotions – 31%
Pricing – 30%
Demand Pricing – 30%
Customer Service – 26%
Quality Improvement – 25%
*Based on 167 respondents who have implemented predictive analytics solutions. Respondents could select multiple answers.
13. Predictive Analytics can help to:
– Identify – the casino's most valuable patrons.
– Predict a patron's future worth and/or his or her future
behavior.
– Plan the timing and placement of advertising campaigns.
– Create personalized advertisements.
– Define which market segments are growing most rapidly.
– Segment patrons into groups based on their behaviors and
then create marketing campaigns to exploit those
behaviors.
14. Predictive Analytics can help to:
– Determine a patron's level of gambling skill.
– Identify patrons who come together.
– Identify the likelihood a patron will respond to an offer.
– Identify the offer(s) to which patrons are most likely to
respond to.
– Predict when a patron is likely to return.
15. Predictive Analytics
In their article “Knowing What to Sell, When, and to
Whom,” authors V. Kumar, R. Venkatesan, and W.
Reinartz (2006) showed how, by simply understanding and
tweaking behavioral patterns, they could increase the hit
rate for offers and promotions to consumers, which then
had an immediate – and substantial – impact on revenues.
By applying statistical models based on the work of Nobel
prize-winning economist Daniel McFadden, researchers
accurately predicted not only a specific person’s
purchasing habits, but also the specific time of the
purchase to an accuracy of 80% (Venkatesan and
Reinartz, 2006).
16. Predictive Analytics
The potential to market to an individual when he or she is
primed to accept the advertising is advantageous for both
parties involved; marketers don’t waste time advertising to
consumers when they aren’t primed to accept the
advertisements, but do market to consumers when and
where they might want to use the advertisements.
17. Predictive Analytics can help to:
With predictive analytics, gaming
organizations can easily segment
their customers and coordinate
marketing campaigns to effectively
target each segment across each
outbound channel. A profit curve
(shown in the following diagram)
estimates the profit a casino operator
will receive from a campaign guided
by predictive analytics, depending on
how many prospects are contacted.
The profit this curve predicts depends
on the ranking of the casino's
customers given by a predictive
model, the cost per contact (e.g.,
printing and mailing costs) and the
average profit per respondent.
18. Manipulating Customer Behavior
Successful marketing is about reaching a consumer with an
interesting offer when he or she is primed to accept it.
Knowing what might interest a patron is half the battle to making
a sale and this is where customer intelligence and predictive
analytics comes in.
Customer analytics has evolved from simply reporting customer
behavior to segmenting customers based on their profitability to
predicting that profitability, to improving those predictions, to
actually manipulating customer behavior with target-specific
promotional offers and marketing campaigns.
19. Predictive Analytics – Survival or Duration Analysis:
A branch of statistics involves the modeling of time to event data;
in this context, death or failure is considered an 'event' in the
survival analysis literature – traditionally only a single event
occurs, after which the organism or mechanism is dead or
broken. Survival Analysis is the study of lifetimes and their
distributions. It usually involves one or more of the following
objectives:
– To explore the behavior of the distribution of a lifetime.
– To model the distribution of a lifetime.
– To test for differences between the distributions of two or more
lifetimes.
– To model the impact of one or more explanatory variables on a
lifetime distribution.
20. Predictive Analytics – Regression Analysis:
Regression analysis is the process of predicting the continuous
dependent variable from a number of independent variables. It
attempts to find a function which models the data with the least
error. Regression analysis can be used on data which is either
continuous or dichotomous, but cannot be used to determine a
causal relationship. Regression analysis focuses on establishing
a mathematical equation as a model to represent the interactions
between the different variables under consideration. Regression
models are particularly effective to find patron worth because the
model can be used to score historical data to predict an unknown
outcome.
21. Predictive Analytics – Linear Regression:
These analyze the relationship between the response or
dependent variable and a set of independent or
predictor variables. This relationship is expressed as an
equation that predicts the response variable as a linear
function of the parameters. These parameters are
adjusted so that a measure of fit is optimized. Much of
the effort in model fitting is focused on minimizing the
size of the residual, as well as ensuring that it is
randomly distributed with respect to the model
predictions. An important assumption of regression
analysis is linearity, which defines a straight line
relationship between Independent variables and
dependent variables.
22. Linear Regression
As per the attached
graph, we can make the
assessment that an
increase in average bet
also increases actual
win and, using the
straight line, we could
predict how much the
actual win would be
affected.
23. Predictive Analytics – Linear Regression
For the casino and hospitality industry, regression
models can be used to predict a patron's future worth
(Sutton, 2011). Multiple regression models “utilize a
variety of predictors and the relationships between those
predictors to predict future worth” states Sutton (2011).
As an example, Sutton (2011) explains that “a model
built to predict future gaming trip worth might be
generated based on historical information about
theoretical win, actual win, credit line, time on device,
nights stayed, and average bet.”
24. Predictive Analytics – Neural Networks
Artificial Neural Networks (ANN) or just “Neural
Networks” are non-linear statistical data modeling tools
that are used when the exact nature of a relationship
between input and output is unknown.
Neural networks can be used to find patterns in data. A
key feature of neural networks is that they learn the
relationship between inputs and output through training.
25. Neural Networks
Neural networks can be used to
classify a consumer's spending
pattern, analyze a new product,
identify a patron's characteristics
as well as forecast sales (Singh
and Chauhan, 2009). The
advantages of neural networks
include high accuracy, high noise
tolerance and ease of use as they
can be updated with fresh data,
which makes them useful for
dynamic environments (Singh and
Chauhan, 2009).
26. Predictive Analytics - A/B Testing
Also known as split testing or bucket testing, A/B testing is a
method of marketing testing by which a baseline control sample
is compared to a variety of single-variable test samples in order
to improve response rates. A classic direct mail tactic, this
method has been recently adopted within the interactive space to
test tactics such as banner ads, emails and landing pages. For
casino marketers, A/B testing is the most effective way to identify
the best available marketing offer. A/B testing involves testing two
different offers against one another in order to identify the offer
that drives the highest response and the most revenue/profit.
28. Predictive Analytics – Decision Trees
Used to identify the strategy that is most likely to reach
a goal. Decision trees are a decision support tool that
use graphs or models of decisions and their possible
consequences, including chance event outcomes,
resource costs, and utility. Decision trees are sequential
partitions of a set of data that maximize the differences
of a dependent variable (response or output variable).
They offer a concise way of defining groups that are
consistent in their attributes, but which vary in terms of
the dependent variable.
29. Predictive Analytics – Decision Trees
For the casino and hospitality
industry, decision trees can be
used “to identify patron
characteristics that can predict
the likelihood of a patron (or
segment of patrons) to abuse
an offer” (Sutton, 2011). Figure
5 shows a decision tree for
responses to a marketing
campaign using age and zip
code as the variables.
30. Predictive Analytics – Time Series Model
A time series is an ordered sequence of values of a variable at
uniformly spaced time intervals. According to the Engineering
Statistics Handbook, time series models can be used to:
-- Obtain an understanding of the underlying forces and
structure that produces an observed data;
-- Fit a model and proceed to forecasting, monitoring or even
feedback and feedforward control.
31. Predictive Analytics – Decision Trees
A Time Series model can be used
to predict or forecast the future
behavior of a variable. These
models account for the fact that
data points taken over time may
have an internal structure (such
as autocorrelation, trend or
seasonal variation) that should be
accounted for. For the casino and
hospitality industry, a Time Series
Analysis can be used to forecast
sales, project yields and
workloads as well as analyze
budgets.
32. Predictive Analytics – Actionable Intelligence
Customer analytics have evolved from simply reporting patron
behavior to segmenting customers based on profitability, to
predicting that profitability, to improving those predictions
(because of the inclusion of new data), to actually manipulating
customer behavior with target-specific promotional offers and
marketing campaigns.
Predictive analytics can graph a customer’s value over time as
well as anticipate that customer’s behavior. From this analysis,
a casino operator can tailor highly specific, laser-focused
marketing campaigns to each customer in the casino’s patron
database. By consolidating the various patron touchpoint
systems throughout the casino property, the casino operator
can create a full view of each patron.
33. Predictive Analytics – Actionable Intelligence
Drawing on data from casino player cards, predictive models
can set budgets and calendars for the casino's gamblers,
calculating their predicted lifetime value in the process. If a
gambler wagers less than usual because they may have
skipped a monthly visit, the casino can intervene with a letter
or phone call offering a free meal, a show ticket or gaming
comps. Without these customer analytics, casino operators
might not notice what could be a slight, almost imperceptible
change in customer behavior that portends problems. For
example, if a long-time customer decides to cash in all their
player card points perhaps it’s because they are dissatisfied
with their last experience at the casino. Predictive analytics
can quickly spot these trends and alert casino management to
the issue so that they can approach the individual to find out if
there is a problem.
34. Predictive Analytics – Actionable Intelligence
This kind of personalized attention can go a long way in
appeasing disgruntled customers, which might be the
difference between retaining or losing them as a customer.
Predictive analytics can glean data from a variety of disparate
sources, including:
-- Data integrated throughout the casino's gaming systems.
-- Feedback information derived from post-visit surveys.
-- Web data mining from customer’s individual online behavior.
Social media websites.
35. Predictive Analytics – Marketing Campaigns
With predictive analytics, gaming organizations can easily segment their
customers and coordinate marketing campaigns to effectively target
each segment across each outbound channel. For example, if a casino
customer is scheduled to receive all of his or her event promotions via
e-mail, the predictive analytics solution will automatically remove him
from concurrent campaigns being run through other channels. This
ensures consistency and also improves customer satisfaction, since the
organization respects the customer’s contact preference and doesn’t
inundate him or her with multiple offers. Moreover, a predictive analytics
solution monitors channel capacity and usage to eliminate overload,
while distributing campaigns equally across the various channels. If one
channel is at risk of overload, the solution automatically shifts the
remainder of a campaign to a different channel to ensure completion.
This enables organizations to maximize the capacity and value of each
channel without resorting to time-consuming manual monitoring.
36. Predictive Analytics – Marketing Campaigns
By utilizing data from past campaigns and measures generated
by the predictive modeling process, casino operators can track
actual campaign responses versus expected campaign
responses, which can often prove wildly divergent.
Additionally, casino operators can generate upper and lower
'control' limits that can be used to automatically alert campaign
managers when a campaign is over or underperforming, letting
them focus on campaigns that specifically require attention.
37. Predictive Analytics – Marketing Campaigns
Sutton (2011) claims that, when it comes to casino patron
analytics, casino operators must seek answers to the following
questions:
How much is a patron worth, how much can we expect a patron
to lose in the future, and who are the most valuable patrons?
-- What patrons come together?
-- What patrons are most likely to abuse an offer?
-- What patrons are the most and least likely to respond to an
offer?
-- Which offers perform the best?
38. Predictive Analytics – Patron Worth
Once patron worth has been defined, data mining and modeling
techniques can be used to estimate predicted worth in the future
(Sutton, 2011). “Simple metrics based on historical behavior, such
as Average Daily Theoretical Loss or Average Trip Theoretical
Loss, will produce fairly accurate predictions of future worth,”
Sutton (2011) notes. “However, advanced predictive models are
able to predict worth with more accuracy and power by
accounting for both patterns in behavior over time and
relationships between predictive inputs that exist within casino
data,” Sutton (2011) argues
39. Predictive Analytics for the Gaming Industry
Casino operators should keep in mind that data mining will
only be successful if their casino patrons are willing to provide
information on themselves. Privacy is a big issue and will
always remain so in the mobile age. Casino properties that can
honor a patron's privacy demands will find patron loyalty
comes with voluminous amounts of priceless patron data. This
is data that can be used to create marketing campaigns that
should prove highly effective. By understanding what type of
patron is on its property, why they are there, and what they
like to do while they are there, casino operators can
individualize their marketing campaigns so that these
campaigns are more effective than normal campaigns, thereby
increasing the casino property's ROI.
40. Predictive Analytics in the Casino
and Gaming Industry
Other chapters include:
• Customer Relationship Management
• Casino Marketing
• Mobile-izing your Marketing
• Social Media
• Table Games Revenue Management
• The Asian Gambler
• Compliance
• A Winning Solution
Book is available at
Amazon.com
41. Predictive Analytics in the Casino
and Gaming Industry
References:
Chung, H. M. & Gray, P. 1999. Data mining. Journal of Management Information
Systems, 16(1), 11-13.
Kreuze, D. 2001. Debugging hospitals. Technology Review, 104(2), 32.
Kumar, V., Raj Venkatesan and Werner Reinartz (2006). “Knowing what to sell when
and to whom,” Harvard Business Review, 84 (3), 131.
Ramakrishnan, Ramya and Madure, Rajashekharappa (2008). Predictive Analytics:
Extending BI Structure. Information Management. December 16, 2008.
Singh, Dr. Yashpal, Chauhan, Alok S., Neural Networks in Data Mining. Journal of
Theoretical and Applied Information Technology. 2005 – 2009.
Sutton, Scott. 2011. Patron analytics in the casino and gaming industry: how the
house always wins. Paper 379-2011. SAS Global Forum 2011.