Companies grow top line revenue either by adding new customers or by selling more to their existing customers. Customer acquisition is a systematic, well-understood process bolstered with a technology assist through marketing automation software from vendors such as Eloqua, Marketo, Silverpop and Pardot. Raising revenue by selling more to existing customers is not nearly as well understood, with a variety of practices and an even wider variety of results.

For most companies, the overwhelming percentage of their revenue comes from existing customers.  Since this aspect of the business is so important, it’s natural to ask why it hasn’t received the same level of attention as customer acquisition, and why it is less well understood. Here are some answers.

First, existing customers are already generating revenue, lots of it, so companies sometimes have less urgency around this part of their business. Powerful, widely used techniques like direct marketing to existing customers produce steady sales.

Second, sales to existing customers fall into two very different categories, up-sell and cross-sell.  Up-sell happens when a customer buys more of what he or she has previously purchased.  Cross-sell is when the purchase is in a completely different product category. The up-sell part is working in most companies, and it accounts for most of the revenue from existing customers.

Why Cross-sell Is Broken

However from our perspective and experience, cross-sell is ineffective and often completely broken. Yet cross-sell is exactly the way companies can get more revenue from existing customers! An important reason why sales to existing customers gets less attention than customer acquisition is that companies don’t know how to do cross-sell in a systematic and effective way.  Instead they focus on what they do know, customer acquisition.

When you look at how companies market to their existing customers, it is easy to see why cross-sell is broken. Cross-sell is all about which products to offer to customers. However the most commonly used analytic technique, Recency-Frequency-Monetary value (RFM), nowhere considers what products are being purchased. It’s impossible to do cross-sell if RFM is your analytic methodology.

Another familiar approach is logistic regression.  Statisticians create a profile of a buyer for a particular product, which is then used to find customers who match the profile.  In a limited way logistic regression can help with cross-sell. The problem here is that logistic regression doesn’t scale. If you have hundreds or thousands of SKUs, the effort involved is usually insurmountable even with fast computers.  Response models, a variant of logistic regression, are generally inappropriate for cross-sell.

The situation is far from hopeless. While cross-sell is much harder than up-sell, it is still easier than customer acquisition.  You know who the customer is. You know a lot about their past purchasing behavior.  Properly done cross-sell analytics can accurately predict what products a customer will likely buy that have not been previously purchased, and can do it at scale. Relevant cross-sell offers, with each customer getting unique recommendations based on these analytics, are producing spectacular results. You can have both — new customers and new business from your existing customers.

Tags :