Wednesday, August 21, 2013

Getting Beyond the Buzz: Making Big Data Work For You

We’ve all heard the buzz about Big Data and its promise. Specific to market research, the discussion largely centers on whether Big Data will replace traditional qualitative methods and quantitative surveys as a primary approach to developing strategic insights. Marketers often favor Big Data as winner-takes-all in this debate; while market researchers frequently argue that Big Data provides a useful complement to traditional research, rather than being a replacement.

Although the debate is fascinating, I’d like to take a practical approach to help you begin to think about how Big Data could drive strategy and insight at your company:
 
  • What can we learn from Big Data?
  • How can it help us improve business outcomes?
  • Where might Big Data fit into our existing insights program?
  • Let’s start by considering the types of questions Big Data and traditional research help us answer and then move on to a series of thought starters to help spark ideas about how to leverage your company’s Big Data exhaust.
 The What and the Why of Customer Behavior
At this year’s ESOMAR and CASRO conferences, our own Greg Mishkin and Dr. Reg Baker addressed the roles Big Data and traditional research play in understanding the customer experience and providing intelligence to drive business decisions. To paraphrase:

It turns out that Big Data (and behavioral observation in general) are really useful when it comes to describing what people do. But they cannot effectively determine why people do it. The why is best identified by talking to customers and asking them questions. That’s the job of traditional qualitative and quantitative research.

The magic happens when Big Data are combined with traditional market research, allowing attitudes and motivations to be projected onto a large population. Optimizing the interplay of Big Data and traditional research to jointly understand the what and the why is the foundation for Market Strategies’ Continuous Improvement Cycle, which minimizes business risk while maximizing business impact.

Two Plus Two Equals Five
While the notion that the whole is greater than the sum of its parts is cliché, it does tell the story. Combining the behavioral information from Big Data with psychographic and attitudinal information from traditional market research helps companies make more complete, connected and confident business decisions. It allows us to understand why people do what they do. It allows us to understand what can be done to positively influence customer behaviors. And it limits the risks associated with resulting decisions and actions, while maximizing the impact to the business.

Integration Thought Starters
Each company and industry has a somewhat unique Big Data exhaust. But the general business questions we all seek answers to are fairly similar. I’ve chosen three strategic issues to frame our thought starter discussion:
  • Customer Churn
  • Call Center Monitoring
  • Market Segmentation
 Although these issues are merely the tip of the iceberg in terms of the types of insights Big Data can help us uncover, they provide a useful starting point to get the conversation going.

As you read on, keep in mind that we are talking about combining Big Data with survey research. We’re looking at the what together with the why. We’re jointly deriving insight from both behaviors and attitudes. And we’re talking about using projective techniques so that we can marry survey insights from a sample of customers with a larger Big Data source that includes all customers (and sometimes prospects) of interest.

Customer Churn
Brands do many things to assess customer churn: Company records are examined to analyze the churn rate. Surveys are sent to lost customers in post-hoc efforts to understand why they departed. Focus groups or in-depth interviews are conducted among at-risk customers to uncover the emotional forces behind their intentions to leave.

These are all valuable approaches. But suppose we take a combined approach. What more is possible if, over time, we project customer satisfaction measures and churn diagnostics (our traditional data source) onto our customer database (our Big Data source)? Here are some of the issues we could explore by leveraging Market Strategies’ Continuous Improvement Cycle approach:

  • Churn Lifecycle—At what point in the relationship does dissatisfaction typically occur? Does it emerge at two days, two months or two years? Does dissatisfaction taper off or remain steady over time? How long after dissatisfaction surfaces does churn occur? What is the window for intervention?
  • Churn Contagion—When dissatisfaction occurs, is it isolated to individuals, or can it spread to an unhappy customer’s network? Can we identify behavioral or contextual triggers that are related to widespread dissatisfaction and churn among customers with common experiences or contexts?
  • Churn Pathways—Is a single behavioral, contextual or attitudinal trigger driving a majority of dissatisfaction and churn? Or are there churn pathways involving combinations of experiences and attitudes that need to be examined as a process to address dissatisfaction? Where is the threshold in the experience pathway? How can we best test and monitor interventions for maximum impact?

 
Call Center Monitoring
We’ve all had the experience of calling customer service and hearing a recording that says, “This call may be monitored for quality assurance.” Well, it turns out that those recordings are a source of Big Data. What’s really special about this type of Big Data is that it uniquely contains information about the social exchange between your brand (via the service rep) and the customer.

As with churn, there are a variety of traditional approaches to monitoring call center performance. These methods usually involve fielding a survey within a short window of the interaction that asks customers to evaluate the performance of the service rep; whether their issue was resolved to their satisfaction and how likely they are to recommend the brand as a result of their transaction experience. These efforts can be massive, often representing the largest chunk of a brand’s total research budget.

To address our clients’ interests in gleaning incremental insights, Market Strategies developed a Calibrated Monitoring approach that combines expert ratings of call center recordings (the Big Data source) with subsequent survey ratings from a sample of the same interactions from customers (the traditional data source). This approach allows us to explore answers to questions that we couldn’t ask using either approach alone, such as:

  • What is the impact of representative accent or dialect on the customer’s willingness to recommend our brand?
  • Are our representatives showing appropriate empathy, and how is the display of empathy related to customer satisfaction?
  • As a result of the representative’s interaction with the customer, did the customer’s demeanor improve, stay the same or remain unchanged from the beginning to the end of the call?

Now consider what we could do if we append the call detail records, or CDRs, to these data and watch to see if the service interaction (and any subsequent intervention) is associated with positive or negative customer behaviors as time moves forward.

Market Segmentation
Market segmentation is one of the most exciting areas where Big Data can help brands minimize risk and maximize impact. Market Strategies’ blended approach to creating segments has always emphasized the combination of behaviors, attitudes and other variables as a genesis for developing segmentation stories that are strategic and actionable. The emergence of Big Data provides us with a richer array of options for incorporating behavioral and channel marketing linkages into our segment views.

Below are two ideas that illustrate how Big Data could help you better understand and frame your market:
  • Consumer Population Data—Imagine a segmentation approach that uses both Census data (our Big Data source) and customer needs and preferences (our traditional source) as a foundation for creating segments. From a strategic standpoint, such an approach could be useful to assess regional candidates for footprint expansion or to drive the development of needs-based products that vary by geographic factors. From an action standpoint, this approach could identify pockets of customers that are most likely candidates to receive tailored messaging. And these customer pockets could be split into test/control groups to pilot an examination of the effectiveness of different marketing treatments.
  • The Internet of Things—What could be done by incorporating telematics data into your segmentation story? Consider data generated from health monitoring applications on smartphones; from vehicles as they are driven; or from cable boxes as consumers watch programming or flip through channels. Each of these data streams contains information on product usage which, when combined with attitudes and customer evaluations of the product, could produce an interesting segmentation that connects needs, usage and satisfaction. Once key segments are identified, in-depth-interviews (another traditional source of data) can be conducted to understand pain points and delighters in the product experience. The insights from this qualitative approach could, in turn, drive the development of improvement programs to address pain and expand delight. A handful of improvement programs could be evaluated among test and control groups to determine their effectiveness, which could be measured by both changes in usage (à la Big Data) and attitudes (à la survey measures like satisfaction).

The combination of Big Data with traditional qualitative methods and quantitative surveys is a critical step toward ensuring your segmentation has the strategic value that’s necessary to guide your brand and the tactical content needed to facilitate segment specific program development and deployment.

Making It Work for Your Company
I hope that some of these ideas have sparked your interest in Big Data and how its marriage to traditional qualitative and quantitative research can deliver a superior level of insight and confidence in decision making.
Have you determined how to balance your approach to insight development across traditional qualitative, quantitative and Big Data approaches? Are you confident you have the right mix and level of connectedness across data types to minimize risk and maximize impact for your company?


About the Author: Contact Dawn Palace at 734.779.6860 or Greg Mishkin at 404.601.9561 of Market Strategies to learn how we can help your company understand the what and the why of customer experience to make more confident and connected business decisions. Greg will be presenting "Traditional Market Research and Big Data Integration: A case study in what works and what doesn't" at TMRE in Nashville October 21-23.
Enhanced by Zemanta

No comments: