James Keyes: Data on the Fly

 
 
By Baselinemag  |  Posted 2005-08-04
 
 
 

For 7-Eleven CEO James Keyes, data is destiny.

Understanding data allows the convenience store chain not only to know what customers are buying now, but anticipate products they may buy in the future.

So, when women are buying beef jerky and pork rinds, 7-Eleven can determine the reason—they are trying to follow low-carbohydrate diets—and then convince a supplier such as Met-Rx to create low-carb products that will jump off shelves. Similarly, the chain can recognize that customers buy fewer pricey tobacco products when gas prices spike, and adjust its stock accordingly.

Moves like those added up to earnings of $96.5 million on revenue of $12.2 billion last year, up from $62 million on revenue of $10.8 billion the previous year.

7-Eleven's technology is nothing fancy—a point-of-sale system that feeds sales data to employees using desktop PCs. That creates a "human intelligence system" that spots trends and creates products to capitalize on them.

Baseline executive editor John McCormick recently sat down with Keyes to discuss how he manages this intelligence system.

How Does Information Technology Affect Convenience Retailing?

Fundamentally, retailing is about keeping up with consumer change. If I take you back to the origins of 7-Eleven 78 years ago, that was a simple proposition. It was one retailer and one customer. We could talk about what you want to buy tomorrow. Over the past 78 years, the complexity across almost 30,000 stores has made it impossible to speak directly with a customer. With technology, we can understand more about actual customer behavior.

Title: Chief Executive Officer

Key Decision: Build a data management system that supports his natural curiosity and leads to new product sales.

Bottom line: Revenue for 2004 surged 13% to $12.2 billion. Earnings for the year up 55%. Share price near 52-week high.

How Do You Find Out What the Customer Wants? a Lot of Companies Do This Through Surveys, but 7-Eleven Doesn't.

Exactly. In fact, we're sometimes criticized for a lack of market research. Many manufacturers will use focus groups and extensive consumer research to determine what the customer may want. The downside of that approach versus the more empirical approach we're using is that what the customer tells you is often different from the way the customer will behave.

We're able to track actual behavior by having technology facilitate a trial-and-error process of market research. For example, when we're going to introduce a new bottled water, we can have two options. The traditional way would be to do some market research on what customers like in terms of enhancements—flavorings, vitamins added, etc.

What we do instead is try three or four [water products] with different flavorings, with different nutritional enhancements, and we can within a very brief period of time determine how the customer behaved by the actual velocity of the items in different stores. By measuring our sales velocity in different stores with different demographic presence, we can get a read on actual behavior rather than predictive behavior.

What Actual Product Information Do You Look at and How Do You View It?

While a lot of CEOs spend their day looking at spreadsheets, I'm looking at a bottle of beer [and how it's packaged]. Certainly, the data is a driver of our business, but the key to success is being able to anticipate where consumer trends are heading, and sometimes those trends are less apparent in historical data and more apparent in the behavioral trends. We have an advantage at 7-Eleven, with stores across 20 different countries, to be able to [identify] trends in other parts of the world.

In Asia, the aluminum bottle was popular. We were able to use our data to see the popularity of this product in our 7-Eleven stores throughout Japan and China. Because of the confidence in the data, we were able to bring this technology back to the United States, say, partner with Budweiser, and ask them to create an aluminum bottle with the same attributes. It keeps the beer colder, and makes it more attractive packaging, more portable than glass.

And it allows us to partner with Anheuser-Busch, test it in our Florida stores, combine their traditional research with our live test data, and then reach the decision on whether or not we should launch this.

How Do You Spot the One Thing That Says An Initiative Will Work in the U.S.?

There are two levels. One is tactical, which occurs every day at the store level, as operators use data to identify products that are not moving well, and replace them with items that have a better chance of success on the shelf. From a strategic perspective, we're looking at trends, so we've got a very different set of reports that a category team is looking at.

Take the low-carb trend. You wouldn't think of 7-Eleven as a health-food store, but we started to see movement in unusual categories–beef jerky, pork rinds, things like that. We had the data combined with an intellectual curiosity. The technology now helps us satisfy our natural retailing curiosity.

Say we look at data, and pork rinds and beef jerky are running at double-digit sales increases after being flat for 10 years. What's happening? We can start asking the right questions: Who's buying it? Why are they buying it? We learned the female customer is suddenly purchasing these products. We start asking our customers and we discover [they] are buying products that satisfy their low-carb lifestyle.

We then took that data to several manufacturers. One of the first and most successful was the nutritional manufacturer Met-Rx. They had already produced a bar with very high protein. We took the idea to them that if you already have high protein, can you create a similar bar with low carbohydrates? And that could be a more fashionable thing to eat for those on a diet than beef jerky. It turns out they could, they did, and we tested with them a couple of SKUs [stock-keeping units] of this new product. Almost overnight, it took off. It became the best-selling product line in the nutritional bar category, and it expanded the segment from what had been fitness buffs to a much broader demographic.

So How Does That Data Get to You? You Can't Look at Every Line Item in a Store.

The nice thing is, I can. I have access to drill down into a hypothesis. Let's say I see a news report on the availability of a new doughnut that has zero trans-fats and no carbs, and I think, 'Hmmm, this could be an opportunity for 7-Eleven.' I have the ability to drill down into different product categories and understand what is happening in similar products that might have the same characteristics as a doughnut with no trans-fats. And based on data, we can begin to formulate a hypothesis about what a similar product would do, what chances of success it might have.

Is This Something You Like to Do?

I like to dig through the data, but I am driven by curiosity. I'm always wondering what's going on with the customer, and I feed off this idea of change. As customers change, can I discover a product opportunity before my merchandising team does? It's a bit of a game. You know, the last thing they want to hear is, 'Hey, I've got this good idea.' I think it creates a healthy sense of a challenge within a company.

You Have Said That Technology Has Allowed 7-Eleven to Take Its Destiny Back. What Did You Mean?

It's an interesting transition from being at the far end of the supply chain, where most retailers over the last 50 years have been relegated to the role of distributing products. In the old days, someone else made the decisions about what goes on the shelf and where it goes, and basically [our role was limited to] ringing up sales. With technology, we're able to reverse that model where the decision about what goes on the shelves is no longer the purview of the manufacturer or the supplier. It's now our decision. We've literally taken it back. By optimizing shelf space, we maximize the utility of that shelf space, maybe add a third competing product, provide something new for the customer, and increase unit sales per square foot.

You Can Do That Because You Have Data?

Exactly. A supplier makes the decision based on its own market share desires—its objective is to capture as much shelf space as possible, which will increase its unit sales, not ours. Our objective is our own sales per square foot, which may mean fewer spacings from one manufacturer to allow more facings of a more popular product. The net effect of us controlling that shelf space is an increase in sales, and it has contributed to our 34 consecutive quarters of improved same-store sales.