“What is the next ‘best’ product or service we should be offering each of our customers?” This is a challenging question, but the good news is, if you’re asking it, you’re already pointed in the right direction.
It’s typical for retailers to be organized around product groups, so the more common question they ask is, “To whom can I try to sell a pair of shoes (or tie or suit)?” There are a couple of key problems with this approach:
- First, this product-centered approach often leads to uncoordinated, inefficient communications with customers
- Second, it doesn’t consider what the individual customer needs or wants
The net result is a confusing or disappointing customer experience, to which customers are less likely to respond positively.
The Next Best Offer strategy, on the other hand, starts with knowledge about the individual customer; this will help you determine the right product or service to offer next. If done well, this is a win-win approach: Customers get the products and services they want, and the retailer develops a deeper, more valuable relationship with the customer.
There are four key steps to developing an effective Next Best Offer strategy in the retail industry:
- Develop an integrated customer view
Big Data may sound impersonal or even scary, but good information about your customers is foundational to providing relevant targeted offers. Gather everything you know about your customers:
- Current/previous products and services purchased – e.g., shoes, suits, sewing/craft products, jewelry
- Transactional history – e.g., visits a specific store once a week; orders printer ink once a month
- Demographic data (self-reported or third-party) – e.g., age, income, number/age of children
- Channel preferences and usage – e.g., customer rarely visits retail location, online shopper, receives both mail and email communications, mobile app preference
Don’t worry if you don’t have all the information you want. Start with what you have and add more detail later.
- Build the next best offer algorithm
This can be done in many ways, ranging from a handful of simple business rules to a multilayered combination of predictive models.
Typically, the goal of the algorithm is to maximize the expected value from the customer, which combines their likelihood of accepting the offer with its projected value (to the retailer). Your analytic approach should be based on a few factors:
- How sophisticated is your current approach to cross-sell? You need to walk before you run. Don’t try to overcomplicate things if a simple decision tree will be a big improvement
- How much data do you have? A sophisticated model will not be valuable if you don’t have the customer data to drive it
- What are your operational capabilities for delivering varied offers? Even with the best data and model, a Next Best Offer strategy won’t work if you can’t execute
- Execute: Deliver the offers
It’s critical to think through the operational, technical and organizational implications for delivering targeted offers to each customer across channels. You need to provide a coherent experience for the customer who, for example, receives an offer in the mail and then visits a retail location, or phones the call center for more information.
- Measure the results
The first layer of analysis should look at the number/type of offers and acceptance rate. You’ll then want to peel the onion to understand differences by customer segment and channel. Ultimately, you’ll want to see how the cross-sell efforts impact long-term customer loyalty and value to your business.
The results will guide you toward your next steps: What additional information do you need about each customer? How should you refine the Next Best Offer algorithm to match customers’ needs better? How can operations be improved or more fully leveraged to enable results?
Next Best Offer is not a set-it-and-forget-it strategy. As customer needs and your priorities change, your cross-sell strategy must evolve too, so continuous measurement and refinement are critical.
William has earned an MBA from The University of Chicago Booth School of Business, an MS in statistics from Kansas State University, and an MS in applied mathematics from Southeast University. William and his team provide data and analytic leadership to Catalyst’s clients.