Intelligencia’s work in the casino and hospitality industry includes work with some of the largest casino organizations in the world, including Genting Hong Kong (Star Cruises), Resorts World Casino NY, and The Venetian Macao, Galaxy Macau, and the Logrand Group in Mexico. Our ETL work has helped casinos optimize their ETL runtime by over 1,000%, making patron data more accessible and therefore much more useful. We have implemented complex CX and analytics systems like SAS PVO, SAS Marketing Automation and build several analytical models to track customer churn, as well as built RFM and Table Games Revenue Management and marketing optimization models.

Customer Acquisition

Just like every other business, casino operators are always looking for new customers. With the gaming market becoming more and more competitive and saturated by the day, there is always a constant need to attract new customers. Customer segmentation models can be used to build predictive models that identify key characteristics of attractive customers.

 

Obviously, a casino will have no internal data available on customers they don’t already have on their books, so the analysis becomes a data mining exercise using publicly available input variables. Casinos can then target these customers with a view to attracting those who have the traits that they see in their already valuable customers. The best external data to use would be population census data, linked to the internal customers by a location identifier (such as postcode or mesh block). It is acknowledged that in some jurisdictions robust and accurate census data may not be available so the model would be relying on whatever information the casino records on its customers from a demographic and lifestyle point of view.  

 

This approach becomes a classical data-mining problem, where a pool of independent variables are tested for the strength of association with the response variable. Once the relevant predictors are identified and the characteristics and traits are defined, marketing and acquisition campaigns can be targeted at the population towards these kinds of people. This would be something that looks to predict a metric derived from current/past customers. Such a metric could come from a segmentation model that identified the high value customers that are most attractive to the gaming company.

 

There are several approaches that can be used and once the target has been defined, this allows for a parametric equation to be derived. This equation attempts to predict the characteristics that distinguish the desirable customers from the rest. This model can only use publicly available information (although other casino information might be acceptable) as that is how a potential customer would be identified. Current information that the company would have on hand would be age, nationality, gender, and address. Where available, third party data should be looked at to further enhance the findings. This could be census data that gives an indication of further customer demographics and this enhances the ability to hone in on customer sweet spots. Data from our data broker partners can also be tapped.

Customer churn

Arguably, customer retention is both one of the cornerstones of any CRM system, as well as being the most important component of the customer lifetime value (CLV) framework. There are indications that companies have problems managing customer retention. Casinos are unique from many other industries in that their customers are tied into contracts, but many of the retention metrics relevant for contractual firms are also relevant for non-contractual firms. A simple 0/1 indicator of transaction, and a measure of recency are appropriate for both types of companies. 

Intelligencia proposes the following process to develop and evaluate a single retention campaign:

1.       Identify customers who are at risk of not being retained.

2.       Diagnose why each customer is at risk.

3.       Decide when to target these customers and with what incentive and/or action.

4.       Implement the campaign and evaluate it.

These steps are applicable to both proactive and reactive campaigns. Reactive campaigns are simpler because the firm doesn’t need to identify who is at risk—the customer who calls to cancel self-identifies. ‘Rescue rates’ can readily be calculated to evaluate the program, and subsequent behavior can be monitored. The incentive should be substantial because the company is pretty certain the customer will churn. Reactive campaigns, however, can be challenging because not all customers can be rescued, and, because we’re dealing with human nature here, customers learn that informing the firm about their intentions to churn can be richly rewarded with valuable incentive, which can endanger the long-run sustainability of reactive churn management. 

Proactive campaigns are more challenging starting from the basic task of identifying who is at risk. Balancing the cost of false positives (targeting a customer who has no intention to leave) against false negatives (failing to identify a customer who is truly at risk) requires sophisticated analytics.

To discover who is at risk, a predictive model must be built that identifies customers at risk of not being retained, or in general of generating lower retention metrics. The dependent variable could be 0/1 churn or any measure of retention. Table 1 summarizes variables predictor variables for several different industries, all in contractual settings, but many will be useful for the casino industry. 

Factors

Example

Method

Customer Satisfaction

1. Emotion in emails

1. Logistic, SVM, Random Forests

2. Customer service calls

2. SVM + ALBA

3. Usage trends

3. Logistic, NN, SVM, Genetic

4. Complaints

4. Logistic, NN, SVM, Genetic

5. Previous non-renewal

5. Logistic, SVM, Random Forests

Usage Behavior

1. Usage levels

1. SVM with ALBA

2. Usage levels

2. Logistic, NN, SVM, Genetic

Switching Costs

1. Add-on services

1. Logistic, NN, SVM, Genetic

2. Pricing plan

2. Dec Tree, Naïve Bayes, Logistic, NN, SVM

3. Ease of switching

3. Graphical comparison

Customer Characteristics

1. Psychographic Segment

1. Logistic, NN, SVM, Genetic

2. Demographics

2. Logistic, NN, SVM, Genetic

3. Customer tenure

3. Logistic, Decision Tree

Marketing

1. Mail responders

1. Bagging and Boosting

2. Response to direct mail

2. Logistic, SVM, Random Forests

3. Previous marketing campaigns

3. Decision rules

4. Acquisition method

4. Probit

5. Acquisition channel

5. Logistic

Social Connectivity

1. Neighbor churn

1. Hazard

2. Social network connections

2. Random Forests, Bayesian Networks

3. Social embeddedness

3. Decision rules

4. Neighbor/connections usage

4. Logistic

Table 1: Predictors of Churn in Contractual Settings

Source: Ascarza, Neslin, Netzer, Lemmens, Aurelie1

The main goal of a retention program is obviously to prevent churn, therefore understanding the causes of such churn behavior is imperative if you are to design an effective retention program. To identify the potential causes of churn for an individual customer, the variables or combinations of variables that are both viable causes and for which the customer exhibits a risky behavior must be discovered. A competing risk hazard model could be used to predict which of the possible reasons of churn are most likely to cause churn at any point in time. Once the causes of churn are identified, the casino needs to isolate those that are are controllable and those who are not. Both correlates of low retention and also the causes of it need to be identified.

To ensure customer retention is front and center, casinos should be scoring their databases on a regular basis in order to understand the likelihood of a customer churning from their venue. This kind of modeling is prevalent in the telecommunications, finance, and utilities industries, and should be utilized in the gaming industry as well. While a slightly different set up due to those industries mostly having their customers locked into contracts, gaming companies need to stay ahead of the game in retaining their customers.

Anecdotal evidence collected in our discussions with gaming companies have indicated a tendency to ignore customers until they have not been seen for up to two years. At this stage, there might be a marketing activity targeted at the customer for up to 12 months. It could be proffered that, by this stage, it is too late to win the customer back; the customer has probably already made up his or her mind and, once a decision like that has been made, it is almost impossible to reverse it, no matter how attractive any competing offer might be.

One of the hardest parts for a gaming company to determine—as opposed to commercial entities that have their customers on contract and definitely know they are tied down—is whether the customer has categorically churned. It may be that a change in location, circumstances or something else has caused a customer to disappear from the sports book, with every intention of returning. However, statistical measures could be used to identify customer’s whose behavior has changed and the change wouldn’t be attributed to chance.

Historical internal data can be used to model the difference between a churned customer and one who is still engaged. There would be significant metrics in the data that identify the likelihood of churning. Similar to the acquisition model described above, a parametric equation could be constructed that elicits the association and relationship between the target variable and the predictors.

This model would serve as an early warning system for the sports book. It would also be a strategic tool useful to predict whether a customer was deemed worth retaining or not. The model should be run on a regular basis across the entire customer database to understand which customers have reached or are reaching a critical value in their churn score. The theory: these customers would then be targeted with an offer to return to the sports book, in the process avoiding the likelihood of them churning. Alternatively, if the customer is deemed to be of little or no value, there would be no offer forthcoming to entice them to return.


[i] Ascarza, Neslin, Netzer & Lemmens, A. (2017). In pursuit of enhanced customer retention management: Review, key issues, and future directions. Customer needs and solutions, 1-17. DOI: 10.1007%2Fs40547-017-0080-0#citeas.

Customer conversion model

The Customer Conversion model can be used to score customers based on information contained in the casino’s source systems as it would only be applicable for customers who had pre-booked their room (as opposed to walk-in customers). Historical information would be extracted from the casino’s IT systems around desirable customers. This would include spending patterns and profitability.nTo identify the relationships that may exist between how the customer comes to the casino and his or her desirability metric, information would be extracted from the casino’s source systems. For a casino, this would include information such as source of betting, channel of betting, lead-time for betting and the incentives offered to attract the customer. Basically, anything that can be attributed to the initial transaction the customer has with the casino would be used as a potential input.

 

These models might also have to be stratified by itinerary to identify the most relevant relationships. The major advantage of a predictive model with this intention would be that it allows the casino to identify customers that they need to interact with once they step onto the casino floor. This would give the casino hosts the potential to get the required information they need to successfully foster a strong and lasting customer relationship.

 

Furthermore, if every potential customer has a score associated with him or her as to his or her long-term likelihood of being attractive, the casino can further hone in on its customers by monitoring their behavior once they are on the casino floor. It is imperative that the casino interact with desirable customers before they have left the property. If customers are made to feel like they are valuable and worthwhile, the likelihood of them returning under their own volition significantly increases. a baseline for customer ROI get also be set at this time, something that can help with marketing expense as the relationship grows.  

Customer segmentation

A customer segmentation model provides a view of the casino from a customer perspective: such models have many and varied applications. Customers are segmented according to what they present to the casino. Views include:

  1. Game preference
  2. Day of week
  3. Time of day
  4. Length of session
  5. Size of stake

Generally, the data is used to determine the appropriate segments for these views. However, the casino has the ability to select the intervals that are preferential and relevant to their venue. For example, it may be desired to split time of day into three, eight-hour periods or six, four-hour periods.

 

The results of this analysis presents a detailed view of how the casino is populated at different times and can allow for appropriate strategic decisions to be made. These decisions could be a function of marketing, operations, or strategy. The output is also used for the building of acquisition models as discussed below.

 

Other potential for analysis would be a master segmentation model that uses the preference results described. Customers are clustered based on their preferences to gain a global view of the casino that is concise and understandable. Furthermore, such models can help measure the impact of strategic decisions, e.g. the addition or removal of a game can be measured against how particular metrics are affected.

The analytics powerhouse SAS is finding its vaunted place atop the analytics pyramid challenged, not just by their typical acronymed competitors — SAP, IBM, EMC, HDS, and the like — but also by the simpler visualization toolmakers like Tableau, Qlik, and Alteryx, who are muscling their way into the mix. These companies offer tools that include data blending, associative engines, and in-memory technology that allows business users to access complete data sets at the touch of a button. These companies offer less complex analytical solutions, but such things as market basket analysis or simple decision tree networks can be created with them and the costs associated with them can be one quarter or one fifth of what the top echelon analytics providers charge.

Intelligencia can help its clients understand the rapidly evolving analytics market, so that they buy just the right amount of analytics for their unique needs. ROI must be considered when it comes to purchasing analytics and we can show you what's available in the market, as well as what should be avoided. Our clients call us 'trusted advisors' and we wear that moniker with great pride.  

According to Gartner’s Magic Quandrant for Business Intelligence Plaforms, modern analytics and business intelligence platforms represent mainstream buying, with deployments increasingly cloud-based. Data and analytics leaders are upgrading traditional solutions as well as expanding portfolios with new vendors as the market innovates on ease of use and augmented analytics. This is becoming an incredibly tricky and complicated field, with new players popping up every years and old stalwarts innovating in impressive ways (yes, we're looking at you, Microsoft). Intelligencia has over a decade's worth of experience helping its clients understand the changing BI and analytics landscape. We can help you understand the options available and then help with a quick and painless implementation. 

As Forbes states, "Customer experience can include a lot of elements, but it really boils down to the perception the customer has of your brand. Even if you think your brand and customer experience is one thing, if the customer perceives it as something different, that is what the actual customer experience is."

Customer Experience (CX) solutions include aspects of CRM, loyalty, multi-channel marketing, analytics, and even social media. Implementations are highly complex, taking into account multiple source systems, strong data cleansing tools, detailed loyalty programs that track every dollar and every secret and non-secret loyalty point, as well as marketing systems that both send out digital content and track offers used. Social channels are even becoming important avenues to connect with and even find customers. Underpinning all of this is a strong analytics base that can optimize every marketing dollar spent. 

 

Implementing a strong CX system is not for the faint-of-heart and we can show you the best available options and systems available to you. 

The goal of data integration (DI) could be considered extracting data from operational systems, transforming it, merging it into new datasets, and then delivering it to an integrated data structure built for marketing, analytics, loyalty, and/or social purposes. The variety of data and context for analytics is expanding as emergent environments — such as nonrelational and Hadoop distributions for supporting data discovery, predictive modeling, in-memory DBMSs, logical data warehouse architectures — increasingly become part of the information infrastructure. Data integration challengers are intensifying because of the increased demand to integrate machine data and support Internet of Things (IoT) and digital business ecosystem needs for analytical processes.

Intelligencia has helped casinos, cruise lines, sports books, social gaming, healthcare, and retail companies integrate their data to create singular customer views of their patron databases that help with marketing, analytics, and social media outreach. We can help you understand the radically changing landscape as well as show you how open source tools could augment or even replace parts of your current DI infrastructure for healthy ROI returns. 

Gartner defines the multichannel marketing hub (MMH) as a technology that orchestrates a company's communications with and offers to customer segments across multiple channels. These include websites, mobile, social, direct mail call centers, paid media and email. MMH capabilities also may extend to integrating marketing offers/leads with sales for execution in both B2B and B2C environments.

A MMH solution can enable businesses to develop and manage personalised customer communications strategies and the timely delivery of offers to its patrons. It allows users to rapidly create, modify and manage multi-channel, multi-wave marketing campaigns that integrate easily with any fulfillment channel, automatically producing outbound (contact) and inbound (response) communication history. Users can define target segments, prioritize selection rules, prioritize offers across multiple campaigns and channels, select communication channels, schedule and execute campaigns, and perform advanced analyses to predict and evaluate the success of customer communications.

Multichannel marketing is one of Intelligencia's strengths. We have worked with large US-based casinos, huge IR's in Macau, and large sports book in Australia, as well as other gaming companies throughout Asia to implement powerful multichannel marketing solutions on a multitude of vendor products. When you're emailing to a hundred thousand people a week, you better have a powerful system and strong support. We'll help you choose the former and the provide you 24/7 of the latter. 

A company is only as strong as its weakest customer relationship and we believe that social media can help Emerald reach their customers in highly efficient and what can be extraordinarily affordable ways. Social media can help businesses in a multitude of ways, including:

  1. Adding interactivity to its Website
  2. Brand and Anti-Brand management
  3. Brand loyalty enhancement
  4. Building fanbases
  5. Crisis management
  6. Discovering important brand trends
  7. Driving traffic to a website
  8. Engaging customers and potential customers
  9. Harvesting customer feedback
  10. Influencer marketing
  11. Reputation management
  12. Social Shopping

Social Media Analytics is the practice of gathering data from blogs and social media websites and analyzing that data to make business decisions. The most common use of social media is to mine customer sentiment. It evolved out of the disciplines Social Network Analysis, Machine Learning, Data Mining, Information Retrieval (IR), and Natural Language Processing (NLP). Social media analytics is a powerful tool for uncovering customer sentiment dispersed across countless online sources. This analysis is often called Social Media Listening or Online Listening. The analytics allow marketers to identify sentiment and identify trends in order to better meet their customer’s needs.

Intelligencia understands the importance of social media and how it has become one of the most important channels to use to connect with today's highly mobile audience. We even understand the Chinese social media markets, so reach out to us if you want to connect with an audience of over a billion people.