Loyalty lost? Your best customers are cheating on you

Many retailers are in the dark about how loyal their best customers are. Yes, they realize that those customers do buy at other stores, but they don’t know how much they spend there.
New findings reveal that many retailers’ best customers (defined here as the top 20 percent of customers by spend) aren’t really all that loyal. In fact, they spend a lot at competitors.
This “lost-sales” issue stems from a flawed assumption: that retailers’ “best customers” shop mostly with them. But that’s what comes from measuring spend dollars only and not tracking true loyalty. As a result, retailers are losing millions because their loyalty spend is being allocated to customers who aren’t loyal, and other marketing spend is not targeting customers effectively.
Retailers are losing millions because their loyalty spend is allocated to customers who aren’t loyal, and other marketing spend isn’t targeting customers effectively.
Our findings, using data from Cardlytics, have  uncovered some uncomfortable truths about loyalty. For a start, it’s apparent that even in segments where loyalty is high, many best customers are quite promiscuous. For example, in do-it-yourself (DIY) home improvement—the most loyal sector in our research—an average of almost a quarter of the total DIY spending by a company’s best customers—actually goes to other home-improvement retailers. The breakdown is almost exactly the same for a leading big box retailer. For a leading upscale apparel department store and quick-serve restaurant, closer to half of spending by a company’s best customers actually goes to other retailers in the category. For every retailer we examined across eight sectors, this promiscuity held true.
Using these new insights can not only shed light on those “lost sales” but can help retailers tune up their loyalty marketing—tailoring their efforts to previously unseen customer segments where sales opportunities are surprisingly large. We believe that two major changes would help:
Best-in-class retailers are going beyond historic spend numbers, using new data and more sophisticated approaches…
1. Measure loyalty, not just spend. In our experience, many retailers have yet to realize the value of true customer spending behavior. Measuring actual loyalty requires coming to grips with share of spend, not just how much customers have spent with you. Best-in-class retailers are going beyond historic spend numbers, using new data and more sophisticated approaches to get better views of their shoppers’ spending habits both in their own stores and with other retailers.
First, they’re using their own customer data to estimate total category spend using analytical approaches such as “twin sister” analyses (Mary “looks” like Beth  in terms of buying habits, except that Beth buys in more categories than Mary does).
Secondly, they’re augmenting their own data with third-party data sources that provide information on actual customer spending habits. One example: They are taking credit and debit transaction data (already collected by many vendors) and combining it with loyalty-card data to understand the customer’s actual spending patterns in a particular retail category.
Are you “leaking” spend with your best customers because of location, format, merchandising, pricing, or promotions?
The results of this kind of analysis are often surprising. For instance, we see that the largest opportunity to capture lost spend is not with sporadic shoppers (as is often assumed) but with best customers—that top 20 percent of customers by spend.  The Cardlytics data on that same large DIY chain discussed above reveals that fully 45 percent of all uncaptured spend is concentrated in the retailer’s best customers That statistic is absolutely in line with McKinsey’s experience across an array of retail sectors.
But it’s not that all best customers are particularly disloyal. If we dig further, we see there’s a subset of customers who are not only a particular retailer’s best customers but also its competitors’ best customers. The statistics on the DIY chain underscore what we see as typical throughout retail: Approximately two-thirds of best customers are truly loyal (defined as giving more than 75 percent of their DIY spend to that particular retailer), while another 20 percent are somewhat loyal (spending 51 percent to 74 percent at one retailer), and 10 percent are what we call “disloyal” (meaning they give less than 50 percent of their spend to the retailer).
2. Discover why those customers are shopping elsewhere—and learn how to encourage them to shop with you. Once you can gauge actual customer loyalty and see where the missing spend opportunities are, the next step is to understand why you’re “leaking” spend, particularly with your best customers. Is it because of location, format, merchandising, pricing, or promotions?  
Behavioral analytics can help here, enabling you to mine the data on “disloyal” best customers (the ones who spend plenty in the category, but less with you  so you can spot, for example, basket holes or missed trip occasions. For instance, segmenting by category behaviors, zip code/location, demographics, or other variables can generate insights into the differences between loyalists and occasionals. These insights can be supplemented with traditional research—going out and actually talking to these customers, for example!—to provide deeper insight and context.
Then it’s possible to develop customer strategies to directly pursue the “disloyal” best customers. These could involve targeted marketing programs—offering personalized product coupons to drive spend into new categories, for instance, or offers intended to build larger baskets per trip. Or you might consider programs that attack the root causes of low loyalty, perhaps by addressing pricing and assortment flaws that cause customers to shop at other retailers, or opening stores in areas with high concentrations of high-spend occasional customers.
When it comes to re-evaluating your best customers, these analytical tools and techniques applied properly to the right dat, can make the difference between inference and insight.