The ‘customer science’ of turning information into retail profit

Is Dunnhumby’s fine-tuned data analysis the way of the retail future, or will it be resisted?


Dunnhumby describes itself as a "customer science company", which is a handy way of describing what you get when you cross big data and fine-tuned data analysis and apply it to the retail sector.

The UK company, whose name is a combination of the surnames of its two founders, Edwina Dunn and Clive Humby, has been a pioneer in the area of customer data analysis for more than a decade. Dunnhumby made its name as the force behind the Tesco Clubcard, one of the first and most successful loyalty-card ventures, a template copied by many retailers since.

Now owned by Tesco, Dunnhumby has been in the news this month after US buyout firm TPG reportedly made a £2 billion bid for the division in early October, a huge increase on the £30 million Tesco handed over for the business in 2001.

Just the week before the buyout news broke, Dunnhumby had been on the keynote stage at Oracle OpenWorld in San Francisco, discussing how it uses Oracle products and services, as well as those from other companies, to provide customer insights to some 400 clients.

READ MORE

In an interview after the keynote event, Dunnhumby chief information officer Yael Cosset said: "Science is the core of our business. Yes, we are very innovative and considered leaders in the industry of how to absorb and interpret data." But the heart of what the company does is figuring out "how to structure your models – it's really about creating the best algorithm and the best models".

This is tightly coupled with how to execute an actual retail strategy around that information. It’s a mix between pure science and the application of that science, he said. “You can’t do one without the other.”

The company’s clients, he said, “are all very committed to providing the most personal service they can to their customers, even if only in their stores, or in and out of store. It’s really about collecting any and all data you can on your retail customers. And also identifying specific trends, product consumption and adoption.

“We help out partners to understand how people want to communicate, and what matters to them. Not all consumers will have the same priorities. Our job is to try and understand what matters to you.”

Customers were like snowflakes, he said. “It’s really easy to see this blanket of white on the ground, but really every snowflake is different.”

Handling the volume of digital data that is now generated by customers and retailers – whether through cards, website use, devices, sensors, or mobile data – is “a massive opportunity, but challenging”.

The value of the insight the company can gather from consumer engagement, such as what customers look for in an online shopping session at Tesco, is really detailed, Cosset said. “Your behaviour online, your geolocation, the weather, may all influence what you eat tonight.” Hence, this can be used to direct customers to an appropriate coupon or offer. Or, as Cosset said, if Tesco knows you bought items in the morning, and that you’ve stopped into Tesco on way home, you shouldn’t receive the same offers.

Personalised services

Wearable devices, such as Jawbone health monitors or the AppleWatch, could provide the type of information that allows retailers and service providers to offer new, personalised services to customers, Cosset believes. These could be recommendations for healthy meals or recipes, or a retailer assembling the ingredients for a meal on a tray, to be collected at the supermarket on the way home.

“It is a massive opportunity, although it is probably going to take a while before we have enough critical mass on specific devices for the data to really be meaningful,” Cosset said. “These are services that consumers are starting to ask for, and the data [from such devices] is extremely valuable in that context.”

Yet surely gathering personal details at this fine-grained level also throws up potential privacy issues, even hovering on the edge of a creepiness factor?

Cosset agrees that privacy concerns will shape how data is generated and used. This will likely be complicated by differences between what various countries or regions, such as the EU, deem acceptable use and management of personal data.

"There are going to be a lot of changes, whether in the US or Europe, from a legislative perspective," he said. "Some things are going to be considered acceptable and some deemed unacceptable. Our philosophy is that as long as you are absolutely transparent with your consumers about the data you are capturing, what you do with the data and what value the consumer ultimately gets from that relationship, then I think you will be in the clear.

“This means articulating [what you do] more than once, in very large capital letters, as opposed to at the bottom of a 150-page document.”

And customers are only going to give up personal information if they see privacy protection and value in return. “If I feel you are treating me fairly, then I will open up the door. If I care about my health, and I am using a Jawbone wristband and you have my transaction history about what I buy and eat, I may say you have been very good at giving me health advice or diet recommendations, and this could mean I am going to give you access to more of my digital data.

“But I think the test is going to be, and is already: what’s in it for the consumer? And basically, I think that what [any data protection] legislation doesn’t catch, the consumer will catch. There will be excessive and abusive uses at some point [from some companies], and the consumer will correct it.”