Dunnhumby on retail's winners and losers

Profile picture for user jreed By Jon Reed April 12, 2018
Can data science make the difference between winning and losing in retail? Dunnhumby's answer is yes - but with a big catch. You have to get the value proposition right. Their Retail Performance Index explains why.

The most powerful storyline from NRF 2018 was the shift from "retailapocalypse" to the reality of retail's winners and losers. Plenty of retailers not named Amazon are in growth mode - but what separates the two groups?

My conversation with dunnhumby in New York City brought some clarity. At the time, dunnhumby had just released their Retail Performance Index. Given dunnhumby's tagline is "Leverage the power of customer data and science to drive sales and margin," I expected to hear a lot of razzle dazzle about AI and data science as differentiators.

Two types of retailers

But that's not how Jose Luiz Gomes, Managing Director North America at dunnhumby, kicked off our conversation. "The new retailing" isn't dictated by tech, but by value proposition. Gomes:

If you don't have the value proposition that's really catered to what your customers want and need, you're going to struggle.

That came out loud and clear in dunnhumby's Retail Performance Index:

What we see in this study is that there is a segment of mainstream retailers where the value perception that they're delivering to customers - and the price that they're charging for it - is out of whack.

Dunnhumby arrived at these conclusions by combining financials and customer perceptions into one report:

What normally happens with these studies is either retailers do really well on financials, but on people's preferences, they do terribly. So we wanted to get the two together to link it up.

Based on this criteria, some retailers are winning:

There are two types of retailers. There are new retailers that open stores where those customers are located. So think the Wegmans of the world who are doing really well. Or Aldi, or Trader Joe's.

Retail targeting - where the data comes in

Other retailers are in a world of hurt and footprint contraction. One major issue: customer segmentation. Customers are spreading retail dollars around in a "fragmented" shopping environment. It's an unforgiving environment if you can't segment properly. And that's where the data comes in:

It's really important for the mainstream retailer to be able to segment their stores at the bare minimum to be able to deliver and cater to the different segments. And what we think is even more important than that is really being clear on the customers that you want to serve, and the ones that you can serve better than your competition.

That means using data to make hard decisions about what you're NOT going to pursue:

The biggest challenge I have with executives is not getting them to understand what they should be doing, it's really prioritizing on what they shouldn't be doing. What do you need to stop doing? You can't be the best at everything so how do you make tough decisions to do that. And that's where the data comes in - it give you the confidence to be able to invest in some areas, and divest in others.

Retailers in growth mode tend to have a targeted advantage. As dunnhumby puts it:

The more targeted a retailer's sites, the more a retailer, like Whole Foods, Sprouts, or Costco, can focus and differentiate their prices, products, and store experience.

Forget the knee jerk imperatives

But here's where it gets interesting. To get the right mix for your customers, you might end up throwing cliches about being "agile" or "real-time" out the window. Even terms like "convenience" or "digital" get called into question. It's about what works for your customers. Gomes cites the surprising example of Costco, where convenience is not necessarily job one:

We saw some of the really strong performing retailers have less positive experience in terms of convenience. Take Costco. Customers appreciate that they're making a trade off between the ease of shop, the prices and the value proposition and convenience. And they're willing to make that happen. So, interestingly, you see that getting that value proposition right will actually overcome some of the barriers to shopping.

(As I type this, I think of a recent visit with my mother who insists on getting all her gas from Costco despite the inconvenience of the location and the gas lines she's willing to wait in. My equivalent would be the ridiculous lines I am willing to wait in to check out of Whole Foods in Manhattan).

"Digital" and "social media" are two more so-called imperatives that might be the wrong knee jerk reaction:

Every time a retailer has to setup a digital offer and a social media hub, the question is: is that really what the customer wants from you, or are you doing it because you feel like there's a market imperative to do it?

Customer science is about asking the right questions

One problem with "winners and losers" thinking: we tag a company with a label, forgetting they can change their fate with better execution. So I asked the dunnhumby team for some client examples. Gomes mentioned Raley's, which has thrived in the California foods market:

Their value proposition is very defined. They're a privately owned company. Mike Teel, the owner, really set the vision, which is they want to transform the grocery industry to help Americans be healthier one plate at a time. So the mission is incredibly strong in what they're trying to do. What we've helped the team do is really segment the way that they approach the market to be able to win. We think Raley's are a real success story and Californian market.

And how does that segmenting work? Gomes used the example of minimally processed cereal:

They've used data from products to create attributes at shelf edge to drive that value proposition of health. So there's things like "minimally processed." They've created some attributes that they think are important for customers. That's using data science to drive their value proposition, which is: "I want you to enable you to make better decisions." So they're making that at the shelf edge easier for their customers to be able to choose when they're in the cereal aisle, I want a minimally processed product.

The wrap - linking data science and value proposition

For Gomes, the power of data science for retailers is taking the flagship store experience, and scaling it up:

We're trying to do that on scale. For million and millions of customers. The customer science, the real power starts by understanding what's important to customers. And then the execution of it: how do you actually get value out of it? Once you've understood what they wanted, then you can use the science to give it to them at scale.

Asking the right questions from the data is the key. The truth is that the customer science conversation and the value proposition conversation can't be separated.

You have to know what you're optimizing for. Too often you don't know what you're optimizing for. What's the best assortment? It depends - who do we want the assortment to work for?

Once that's decided, the hard work of culture change begins:

When people tried to copy the work we did with Kroger, what they saw were the widgets and the software. What they didn't appreciate was the cultural transformation that really went into it.

These types of indexes are provocative, but do have limitations. One year's winner might fade if execution falters, or if a beefed up competitors enters the fray. Example: will an Amazon-powered Whole Foods make life tough on Trader Joe's? Time will tell. But the data implies that your retail fate is in your own hands, not dictated by store footprints or assumptions on "going digital." Let's see how this plays out in future editions of this index.

End note: David Cianco, Senior Customer Strategist for dunnhumby, joined this conversation and provided important context to the quoted material.