Suppose you learned that most of your brand's buyers are switchers... that only 15% of your customers are highly loyal to your brand and account for maybe half of sales. The harsh reality is that that's a typical pattern for many grocery brands, for example.

You're probably thinking, Can this really be true? Isn't marketing about building engagement to create and then "lock in" super-valuable customers? Well, here are some sobering facts about the average brand:

  • Half of your loyal buyers this year will not be loyal to you next year (Catalina Marketing's analysis of tens of millions of shoppers).
  • The 20% of buyers who account for 80% of sales includes super-heavy category users who might even prefer another brand and purchase that brand more.
  • On average, 30% of loyal buyers do not have attitudes about your brand that support their loyalty and are the ones who are most likely to defect (from a paper I coauthored in the Journal of Advertising Research in 1996).
  • For most brands, only a single-digit fraction of your customers connect to you via social media. Consider Coca-Cola on Facebook. When I last looked, Coca-Cola had 3.5 million fans, but the number of people on Facebook who drink Coke is probably north of 100 million.

So how do you build a marketing plan to support the other half of your sales—the half that comes from buyers who are not really loyal to your brand?

For those buyers, it isn't that brands don't have meaning; it's that more than one brand has meaning and is trusted. As a shopper decides what to put into the shopping cart, you are competing for that purchase with other brands that have meaning for that shopper, often right at the point of purchase, with your brand sometimes the preferred brand, and sometimes not.

But are those "occasional" buyers really important? When a brand grows, doesn't it do so by converting low-loyalty buyers into becoming engaged with the brand—so sales from those who remain low-loyalty buyers actually go down?

Let me shake your belief structure a little more. Modeling of longitudinal purchase data proves that you must succeed with people at all levels of loyalty if you are going to grow your brand.

Based on the statistical ("beta" and "Dirichelet") distributions that modelers use to analyze purchases for frequently bought goods, consider what happens when a typical 10%-share brand grows to become a 15%-share brand.

If you segment category buyers by loyalty level into quartiles, you will see that you must grow share in each quartile! The high-loyalty group must go up the most (the Ehrenberg double-jeopardy effect), but even those in the low-loyalty group must increase their purchases of your brand by 30% or so.

Sign up for free to read the full article.

Take the first step (it's free).

Already a registered user? Sign in now.



Joel Rubinson is chief research officer at the Advertising Research Foundation ( and blogs at Reach him via or Twitter: @joelrubinson.