Management Overview
The essential concept of data (big, small, transactional, structured and unstructured, modeled and predictive, and so forth) is neither new nor unique. However, the attention to, and focus on, challenges associated with “big data” has grown to the point where key elements of its relevance—accuracy, quality, analytics, monetary value, and so forth—are being called into question. Companies need to have the right, high-quality customer-related data, which is consolidated from multiple sources into a single view. Then, to build or sustain customer centricity, those data need to be managed, analyzed, and reported to best, and most monetizing, effect.
Example of Big Customer Data Actionability: At-Risk Customers
Customer risk is one of the most important, and most overlooked, elements of the customer life cycle. There are many data sources for identifying potential customer churn; and there are also multiple actionable applications for such information: service and overall experience processes, employee performance, and so forth.
Insights Lost from Dropping Loyalty Programs
As the desire for, and understanding of, customer data continues to increase, there is also evidence of prominent retailers eliminating the loyalty programs which not only reward their best and most active customers but ensure the flow of data for ready evaluation. In sum, the data produced by loyalty programs is not seen as contributing to return on investment (ROI). This is reflective of insufficient analytical capability on the part of the organization eliminating its program.
Costs Associated with Low-Quality Customer Data
With the proliferation of customer data available to marketers, data quality has become an extremely important issue. Data must not only be complete and accurate, and consistent as to source, they must be relevant and easy to use and combine into data streams. The errors can range from duplicate names on files to poor merge–purge software. Many companies are utilizing frontline staff to conduct data quality checks; however, this is inefficient; and the responsibility can be handled more efficiently and less extensively through linkage software.
Correlation Is Not Causation
Organizations assume that statistically valid relationships between one set of data and another mean that they are correlated. This logic flaw applies in all areas of customer centricity: employee satisfaction and engagement relative to customer behavior, and, especially, connections between customer attitudes and behaviors. Rather than rely on correlation, there are advanced forms of predictive analysis which will identify behavior patterns on a micro-segmented customer basis. In addition, more extensive testing of branding, product, and communications concepts is actively encouraged.
“At-Risk” Customers: Do You Have a System for Identifying and Stabilizing Them?+ (September 23, 2013)
About a decade ago, my consulting colleague Jill Griffin and I identified seven distinct customer life stages for our 2001 book, Customer WinBack. These life stages, or components of the life cycle, could be applied to customers of any type, and any size of enterprise. We considered the most serious, and potentially impactful, of these to be customers to be those “at risk.” These customers have a proven high probability for defection. A decade later, that perspective hasn’t changed. Because the average company loses between 20% and 40% of its customers a year, isolating drivers of risk and stabilization (i.e., repairing and rebuilding the relationship) is a priority for any enterprise.
We identified multiple causes of value, or trust, breakdown leading to the stage, or state of risk. It could be a negative transactional outcome with the supplier of the product or service, or a series of them (at least as perceived by the customer), or any of several factors that can impact favorability levels, with the result that the “secure” customer will begin to exhibit specific behaviors and actions that indicate a weakening of the relationship. The sooner intervention begins with an at-risk customer, the easier and less expensive it is to fix the problem; so, it is clear that the best plan is to have a system for identifying these customers. That said, many companies have no such system, or even a basic risk-pulse process, in place.
Here is what we recommended for a first-alert information system that receives input from a variety of “nerve centers” throughout the organization, and also outside. Some of these are reactive, providing insights after the fact. Some are more proactive, helping anticipate trouble spots and problems looming on the horizon that may adversely affect customer loyalty. Some are big data related, and can be uncovered through basic analysis. Some are more individualized and require greater sophistication and sleuthing capabilities. Some are both. These include the following:
1. Purchase and use data—declines in purchase or use activity represent proof that customers are spending elsewhere, and are a defection risk.
2. Listening posts—relatively few customers, whether B2B or B2C, actually complain. But internal (frontline employees) and external (social media, ratings, and so forth) posts can provide useful information.
3. Customer transactional and relationship research—can keep track of perceived performance (such as changes over time) with customers overall, as well as with specific customer groups and individual customers.
4. Accounts receivable—“short payment” and slow payment are tripwires to customer discontent and risk.
5. Customer network—querying, or debriefing, advisory boards and online communities can uncover areas of concern.
6. Churn modeling—advances in data mining software have enabled more companies to apply sophisticated churn algorithms, evaluating the interplay of multiple purchase and behavior variables to uncover defection behavior and proclivity patterns that might otherwise go undetected.
7. Account event milestones—firms like banks, credit card, and insurance companies have discovered that certain milestones—such as renewal dates—can trigger risk and relationship reconsideration.
Perhaps the most leveraging element of both risk and stabilization in customer experience is interaction with employees. Beyond employee engagement and its somewhat tangential influence on customer behavior, employees have the capability to be “ambassadors of value,” as well as conveyors of insight. If they are proactive, involved, and committed on three levels—to the company, to the value proposition of the company’s products and services, and to the customer—then risk can be mitigated and the solid relationship, leading to customer loyalty behavior, can be regained.
And, how can the various data points from this RIS, or Risk Insight System, be analyzed, applied, and leveraged to reduce churn? Answer: In virtually every way that can be conceived, including support staff hiring, training, and motivation; service process modification, improvement, or both; customer support proaction and outreach; messaging and communication; product/service repositioning or design enhancement; array and availability of contact channels; and so forth. Perhaps the most important of these is reshaping the culture around customer centricity and employee ambassadorship.
Seduced and Abandoned Customers…and Lost Insights: Dropping Loyalty Programs and the Data Value They Represent (and Making Excuses for It) (July 16, 2013)
Earlier this month, my customer experience colleague Colin Shaw wrote an excellent, insightful blog about the actual customer value that is, or isn’t, inherent in loyalty programs.1 My response in support of Colin’s perspectives was that, as in virtually any worthwhile relationship, there must be mutually beneficial value which, to the extent possible, is both targeted and personalized. Just having a company’s loyalty card (Figure 4.1), per his post, means relatively little unless customers both understand the program and see reward for themselves through active participation.2
Figure 4.1. A company’s loyalty card means relatively little unless it is beneficial to customers.
I also noted that companies behind the loyalty program, likewise, need to see value for the enterprise (in the forms of new customer attraction, product/service differentiation, “barrier to exit” customer retention, data to leverage, and profitable performance) as well as the customer. We’ve seen with some UK supermarket chains, and more recently Albertson’s, Acme, Shaw’s, and Jewel-Osco (sold by Supervalu to Cerberus Capital Management in March) that there is a growing tendency to abandon the loyalty program—definitely a risky, and potentially very commoditizing, move—in favor of, as Albertson’s has told its customers “offering great prices to everyone.” What happened to personalization and tiered customer benefits? For customers, there’s something approximating “bait and switch,” diminished treatment in this move. So much for targeted, personalized customer value.
In this era of big data generation and leveraging insights to create greater customer value and loyalty behavior, Cerberus has made a strategic decision in the opposite direction. The loyalty programs for these former Supervalu chains were pretty superficial and generic, focusing principally on discounts; and, as in the case of Shaw’s, a major value and experience makeover was deemed more of a priority than endeavoring to mine and apply the data coming from the program.
Greg Girard, marketing strategies and retail analytics program director at IDC, said that shoppers “have been defecting from Shaw’s for fundamental reasons—high prices, dingy stores, and poor customer service being the most pressing.” As Claudia Imhoff, president of Intelligent Solutions, commented: “I don’t think they have the analytical prowess to use the information that’s being given to them, and that’s a shame.”
The lessons here are:
1. that the loyalty program has little or no value for customers if it is non-engaging and represents little benefit, beyond just saving money; and
2. that the loyalty program has little or no value for the company sponsoring it if the data collected are not used to improve customer ROI, and all the organization is doing is throwing resources into something that’s expensive and unproductive.
Claudia Imhoff wisely concluded: “We need to get analytics to a point where it’s easy to use and the cost isn’t prohibitive. And, we have a little bit of education to do with low-profit businesses so that they can see a return on investment quickly and in a tangible way.” For Shaw’s, the priority now is to first solidify the behavior of its customer base, or so Cerberus appears to believe. “Fundamentals are the cake of customer loyalty; personalization is the icing. Cerberus has to bake the cake of customer loyalty from scratch -- and in a hurry,” IDC’s Girard said.
The High Cost of Low-Quality Customer Data—Or—Why Does My Wife (Still) Get (So Many) Duplicate Catalogs?+ (June 9, 2013)
First off, let me admit something: in my household, we do our share of direct response shopping, via Internet and mail catalog. A wrought iron floor lamp from Morocco and a wall clock from France for the kitchen, flexible thumb picks for me to play my baritone ukulele, a cast brass walking stick holder from New Zealand for the foyer…well, you get the idea. As a result, our family, especially my wife, gets a ton of catalogs and other promotional content. Not a week goes by that she doesn’t get two of the same catalog, one addressed to her name when she was previously married and one addressed to her name now. Even though we’ve been married for many years, these catalog companies are obviously using databases that are neither clean nor current.
Clean and current data are but two aspects of data quality. According to the Navesink Consulting Group, data are of high quality “…if they are fit for their intended uses in operations, decision making, and planning.” That means the data
• are free of defects: accessible, accurate, timely, complete, consistent with other sources, and so forth;
• possess desired features: relevant, comprehensive, proper level of detail, easy to read and interpret, and so forth.
Using these criteria, even in this era of sophisticated merge–purge in the name of better, more efficient marketing, no customer database is error-free; and the high level of erroneous data has become a high-cost problem for suppliers and an annoyance for customers. In a recent study, close to 80% of consumers polled said they dislike receiving duplicate pieces of mail for the same promotion. But, names can also be misspelled, or mail pieces can be sent to former residents or the wrong address. High-quality data and the sophisticated statistical techniques to analyze this information are absolutely essential for successful customer programs and processes. Everything is potentially important in customer relationships—historical purchase data, essential demographics, and life style characteristics—so suppliers are becoming increasingly concerned.
The range of estimated errors in most databases is quite broad. I’ve seen percentage figures that go from the low single digits to 40%, and higher. Errors can come as a result of keying errors, misheard names and addresses, or straightforward multiple entries of the same data. They can also arise from poor list merge–purge programs, errors when data streams are integrated, infrequent cleaning, and so on.
Costs associated with poor data quality go well beyond the easily identifiable, and obvious, waste from mailing duplicate merchandise catalogs. They include dealing with customer complaints caused by data errors, and the staff costs involved in checking databases, finding missing data, and fixing incorrect data. Order entry staff, for example, may spend up to 25% of their time performing these tasks. In a company with 40 people on this staff, that would equal 10 effort-years annually. Further, if the loaded cost for each of these staff members is $50,000 per year, the annual cost associated with data errors—in just this one area—would be $500,000!
Lutheran Brotherhood (now Thrivent Financial for Lutherans), a member-owned fraternal benefits organization (mutual funds and annuities, insurance, estate planning, college and retirement programs, and so forth), has several million member names and addresses spread across several product and service line databases. In addition, they have millions of names in their prospect database. Several years ago, they often ran into customer service problems, one of which centered around delivery of their bi-monthly magazine. If a household had multiple members, and they requested multiple copies, there were questions about why they were getting only one copy. Also, like my wife with merchandise catalogs, the same member might get multiple copies of the magazine even if they should have received only one.
The organization was principally using frontline staff to rectify member data quality problems, but this was both expensive and inefficient. They undertook a process to centralize all member information into one database (customer data integration), so that employees would have a complete view of their members. This was a proactive and positive first step, but it didn’t eliminate the bigger problem of duplicate names and addresses. Staff were still responsible for the name file searching and matching process, functions that were costly, time-consuming, and negative for morale. According to a senior business analyst for the organization: “We were spending too much time entering and cleaning duplicates.”
They solved this issue by using advanced linking software to de-duplicate member and prospect names from the organization’s database. Their entire file can now be cleaned for duplicates in just a few hours. Much of the problem surrounding magazine delivery evaporated as a result.
Data quality upgrade efforts such as these are usually transparent to customers, and they should be. The real rewards of higher quality data, however, are improvements in areas of customer relationships such as service and market research, and greater efficiency in marketing and other customer-related processes, making customer relationship management (CRM) and customer experience management (CEM) activities more productive and consistent.
I haven’t even mentioned that the lack of personalization and overall data quality sends many bland and untargeted e-mail promotions our way. This is another major data management, customer experience, and marketing communications effectiveness issue, and it’s extremely important in building and sustaining relationships, as well as controlling costs—but it’s a topic for another day.
Correlation Is Not Causation: Big Data Challenges and Related Truths That Will Impact Business Success+ (February 4, 2013)
For years, social scientists and consultants have warned the corporate world about making too much of correlation analysis, the simple regression technique which shows the relationship between one set of attitudes or behaviors and another. As an example, “The Service Profit Chain,” a model developed by three Harvard professors in the 90s, is generally summarized as happy employees = happy customers = happy shareholders. In other words, at the core of effective employee engagement is the tacit belief that there is a direct relationship or linkage between higher employee satisfaction and customer experience. And, as found by noted customer experience expert Frank Capek, though elevated levels of customer service, and also increased profitability, may result from enhanced employee engagement:
Just because employee satisfaction and engagement are correlated with customer satisfaction doesn’t mean that making employees happier will lead to better customer experience. This is one of those classic traps your college professors warned you about: confusing correlation with causation. I’ve observed that this flaw in logic has led many organizations to invest in trying to make their employees happier in the hope that those happier employees will turn around and deliver a better experience for customers. We’ve just seen too many companies where, at best, more highly engaged employees simply deliver a sub-par experience more enthusiastically.
What is true in the world of employee behavior optimization is, if anything, even more of a fact in the broader landscape of marketing, brand positioning, communication, and customer experience. CMOs, for instance, have relied on correlation as a core analytical approach for connecting basic customer value performance, and the identification of unmet needs, to forecasts of potential sales and profitability. Today’s customer, however, is less patterned and more self-educated, more socially connected, and independent thinking; and this sea change has put a great deal of pressure on the kinds of customer information, and analytics, CMOs are used to handling. There are new streams, and types, of data that CMOs will need to understand; and there are also new analytical approaches—getting to key, causative drivers of customer behavior—that will be required for insights into why customers think what they think, say what they say, and do what they do.
This is, to a great extent, where “big data” comes into play. Marketing has always had some volume of available macro quantitative data, such as customer profile, purchase history, ad hoc research, historical brand and transactional tracking reports, and so forth—for looking at decisions involving target audience, new products and concepts, value proposition, and competitive positioning. That said, marketing management and the cultures in which they function have also historically relied more on the conventional wisdom brought about by their own experience and instincts, creative concepts under consideration, and engaging communication.
Customer data analysis, where applied, has largely been of the straightforward correlation or cross-tab type, for evaluating simple, core business elements. In sum, marketing has been more about supporting big ideas than having objective, insightful information. Now, with the kinds of analytical tools which are emerging, marketers can crunch enormous petabytes and terabytes of data from the sources just identified, plus third-party statistical information exchanges, and public and industry statistics, to create all manner of fascinating correlations (Figure 4.2). The worthy goal is to identify connections, or data stream correlations, between one set of customer information and another, in ever smaller and smaller audience micro-segments so that marketing dollars can be more effectively spent.
Figure 4.2. Analytical tools crunch a large amount of data to create highly actionable insights.
This is where users of such data need to exercise care. Correlation is not causation, and the insights produced by big data analysis tend to be only directional, tentative, and preliminary. Even though big data are more complete, and more available, than ever before, there are often missing elements in databases, plus disruptive, or confounding, factors which can compromise data quality. Another way of saying this is that correlation analysis of big datasets generates results that should be seen as more hypothetical than actual; so there is little assurance that any correlations uncovered by big data will directly influence customer behavior.
Rather than serve as an unchallenged platform for decision making, a better use of the insights spun out by big data analysis, and especially correlation analysis, would be to distill the results into testable value propositions. A marketing, and corporate, culture built on testing and controlled experimentation leads to more financially sound, proof-based, answers which will truly help grow the business. Testing of ideas and concepts certainly isn’t new, but it needs to become an enterprise mantra. Leading companies have already embraced testing and experiments as the fuel for an engine of success.
Section Summary and Perspective
Though multiple streams and sources of product and customer information have been with us for some time, “big data” gets a tremendous amount of attention. This is because big data has become a catch phrase for the sheer volume of information in datasets now available for business intelligence and decision making. Over the past 20 to 30 years, business’s ability to generate transactional, interactional, and observational data has moved from basic structured (in SQL databases), to semi-structured, and now to unstructured data generation and mining. It is estimated that, each day, 2.5 quintillion bytes of data are created (so much that 90% of the information in the world today has been created in the last 2 years alone).
Data can come from traditional sources, such as research and customer purchase records and databases, and new unstructured sources, such as e-mails, posts to social media sites (Twitter processes 400 million tweets every day), digital pictures and videos, cell phone GPS records, and so forth. Cisco has predicted that, in 2013, information flowing over the Internet would reach an annual volume of 667 exabytes, 2.6 million times the amount of information stored in the Library of Congress.
Today, companies are challenged as never before to harness the volume, variety, velocity, and value, of big data as waves of information flow into the enterprise. The opportunity presented is to provide the business with greater marketing and brand-building decision-making flexibility and agility, and to address issues that were previously beyond the scope of most organizations.
So, data represents opportunity—to improve processes and customer experiences, to develop products and services with higher value, and otherwise make monetizing changes that will resonate with customers. One application, for example, is reducing risk leading to potential defection, perhaps the most important stage in the customer life cycle. It’s critically important, however, that all customer-related data be a continuous flow, so that actionability is targeted and in as real time a fashion as possible. As pointed out in this section, elimination of a loyalty program, and the customer data it produces, seriously jeopardizes a company’s ability to take value-driving action. It’s also very important that the generated data be of high quality; otherwise its utility is seriously diminished.
Finally, with respect to analysis and actionability, organizations should have a grasp of the most contemporary, and most useful analytical techniques. Otherwise, even the most complete reservoir of customer data can take an enterprise in the wrong direction.