The Evolution of Hyper-personalization: Part 1

Lessons from transformational Kroger/dunnhumby partnership

April 4, 2016, 04:54 pm By Gary Hawkins, Center for Advancing Retail & Technology (CART)

Editor's Note: This is the first of a two-part series on hyper-personalization in the retail food industry.

Supermarket retailers look to the achievements of the Kroger-dunnhumby partnership with awe: 49 consecutive quarters – more than 12 years – of quarterly same store sales growth, improvements in price perception, strengthening of margins, and more.

Kroger has become a master of strategic marketing personalization, the use of targeted promotions to grow the value of shopper households over time. In a December 2015 investor conference call, Kroger CFO Mike Schlotman, spoke to the third quarter’s 5.4 percent same store sales growth as being driven by an increase in the number of households shopping and an increase in shopping frequency.

He went on to call out that Kroger continues to grow the number of loyal households at a faster rate than the number of total households who shop; in other words, Kroger is continuing to transition casual customers into loyal customers at a critical rate.

While the value of marketing personalization by retailers is beyond doubt, the data consultancy-based approach used by dunnhumby and its competitors is massively expensive and has stymied even significant regional retailers, let alone smaller independents. Few can afford the hundreds of data analysts crunching through Kroger’s data. However, just as the Kroger-dunnhumby partnership has transformed retail marketing, advances in technology threaten to disrupt the consultancy-driven, shopper segment-based model that has become the norm.

The explosive growth of low-cost computer processing is expanding the use of cognitive computing, converging with big data and the cloud to bring new capabilities to retailers. Consumer goods retail provides a fertile field for the rich application of artificial intelligence and machine learning to upend past approaches to shopper centricity. Hyper-personalization is upon us; but first let’s take a look at where we are today and how we got there.

Personalization as We Know It

Kroger’s success has been driven by analytics and insights drawn from the massive store of shopper-identified purchase data the company possesses, along with third-party data it acquires. The company’s analysts have done a masterful job of creating shopper segments built on demographics (income, household size, etc.), insights (shopping trip types, basket size, shopping frequency, etc.) and attributes (pets, babies, seniors, organic, gourmet, etc.).

Kroger then leverages these shopper segments to send targeted promotions to specific households in order to strategically grow shopper value; converting secondary shoppers to loyal shoppers, loyal to premium loyal, and maximizing shopper retention. And this promotion targeting is a massive effort: According to Kroger’s 2012 annual report, a Loyal Customer Mailing (LCM) reached more than 11 million households, and 97 percent of recipients received an individualized set of coupons for the products they like and buy regularly – almost no two were alike.

While one cannot dispute the effectiveness of this approach, it requires an increasing number of analysts creating highly refined segmentations that are combined with rules and algorithms to distill the number of promotions for which a given household may qualify. While valuable, this is a time consuming, laborious process requiring multiple iterations of data processing.

Heading Toward Hyper-personalization

In the Kroger/dunnhumby approach, each shopper household is receiving a different mix of offers, each offer targeted to that shopper because the shopper household belongs to some segment. Contrast Kroger’s approach to marketing personalization with hyper-personalization using the latest sophisticated technologies to target the individual shopper; a segment-of-one. A subtle but important - and immensely powerful - difference.