Analytical Life Stages: Using Loyalty Analytics to Grow from Data Infancy to Adulthood
InfoManagement Direct, March 2004
The numbers are staggering. According to a recent study by the IDC Consulting Group, business analytics generates an average five-year ROI of 431 percent. Analytics based solely on customer strategy, the study reported, generated a median ROI of 55 percent.
These results seem too good to be true, but they're not. There are three ways to improve your bottom line: you can sell better, you can retain better or you can manage your current customers better. Every additional metric falls under one of those three categories.
Within the realm of customer relationship management, those sexy ROI numbers have not gone unnoticed. AMR Research estimates that investments in analytical CRM applications will grow at nearly double the rate of operational CRM systems, with the market expanding to nearly $4.4 billion - 19 percent of the total CRM market - by 2005.
Advertisement
But there is a question that many marketers forget to ask; where are you going to get all of this customer data to analyze? How do you get your customers to raise their hands? And is the mere reporting of historical transactional data enough?
Data from loyalty programs can play a crucial role within a larger CRM strategy. Loyalty program databases are generally the most complete, robust and easily accessible in the enterprise. Customers opt in to trade information for value and allow you to track their behavior with the expectation of mutual gain.
Likewise, loyalty analytics - the art and science of analyzing customer data to identify, maintain and increase the yield from your best customers - can be the centerpiece of your approach to analytical CRM. But where do you begin? How do you know if you can even afford to launch a loyalty program to capture customer data? And once you do capture it, how can you maximize your return from it?
It's a complex problem to solve. There are millions of customers, millions of prospects, probably hundreds of offers you can make and interesting interplays of customer value, customer risk and propensity to buy.
Join us as we trace the growth of an enterprise through loyalty life stages, from a loyalty toddler just getting a handle on customer data to a loyalty adult using sophisticated analytical techniques to segment customers and predict their likelihood of churn. Although we are focusing on loyalty applications, the lessons learned from analytical CRM can be universal, allowing you to evaluate the potential of any marketing initiative - whether or not it's related to your loyalty program. It's a challenging path. But if you're willing to crawl before you learn to walk or run, then it can be a rewarding one.
Figure 1: The Data Toddler - Forecasting Program ROI
Within the realm of loyalty analytics, companies generally fall into three classes. There are data-rich companies that already run loyalty programs and want to better understand how to measure them financially. There are companies who have servers full of customer data but who aren't sure what to do with it. There are companies that possess fragmented data or even no data at all but that perceive its value and would like to forecast how a loyalty program will perform in their business environment. We'll call this first stage of growth the toddler stage.
If your enterprise is in the toddler stage, how can you know if you can support a loyalty program with a compelling value proposition? Your first step down this path is to employ a financial template. This template can be internally generated or supplied by a consultant. Either way, the template will be, at its earliest stages, assumption-driven and must be robust enough to evaluate any marketing initiative. As you proceed, remember that it's just a forecast. The financial model becomes better and tighter if you later validate those assumptions with actual data.
There is no ROI rule of thumb, because it depends on the financial philosophy of your company. But the bottom line is that you want to improve your profitability. Backing it up from there, you want to improve the profitability of your customers.
You might know subjectively or through experience that your customer's average value is, let's say, $500 a month. You have no data to support that premise, but you consider it a reasonable number. But let's say that, instead of plugging the single average customer value of $500 a month into your template, you instead identify a statistical model and place it around that number to create a distribution of customer values between, say, $200 to $700 a month. Absent any data to validate which distribution model they should employ, most analysts assume a bell curve. Now, instead of relying on a single assumption, you can model a program's financial performance with a distribution of outcomes, with the added ability to evaluate the potential risk if your loyalty initiative is unsuccessful.
Most enterprises know how many active customers they have. But what percentage of those transactions can you identify? If you have a million active customers, and if each one of those customers transacts on average twice a month, then you have two million transactions per month. If, however, 50 percent of your transactions are cash based, you can't evaluate those transactions nor will any kind of analytical CRM engine help you to leverage that invisible information.
Your current ability to capture transactions thus limits your information set to those one million identifiable transactions. At this point, we have left the general customer universe and have entered the potential known universe of your proposed loyalty program, which can then theoretically attract and identify those invisible, cash-only customers. How do you then filter out customers who are unlikely to be good candidates for enrollment?
Let's return to our $500 a month average-value customers with our statistical distribution of $200 to $700 per month built around them. If your enterprise averages a 10 percent net margin, then those $200-per-month customers mean $20 a month each to the bottom line. If, in your hypothetical loyalty program, you assume a cost of $30 a month to enroll, communicate to and provide a reasonable value proposition for those $200-per-month customers, then their current contribution to margin is less than what it would take to enroll and engage them as program members. No loyalty program for them.
Page 1 of 3.
