Extracting Business from Intelligence

STORES Media editor Susan Reda moderated a panel discussion entitled

“Exploiting Advanced Optimization Capabilities in Core Retail Functions.” Panelists included Karen Etzkorn, senior vice president and CIO of Ascena Retail Group; Herman Nell, vice president and CIO of PETCO Animal Supplies; and Janet Sherlock, vice president and CIO of Carter’s.

In her introduction, Reda said that the general topic of the discussion was business intelligence and the analytics solutions that provide the power to organize, analyze and visualize the terabytes of data that almost every retailer now amasses. What follows is a synopsis of that panel discussion in Q&A form.

By way of initial background, Carter’s is a $2 billion-plus retail/wholesale vertical company utilizing MicroStrategy, QuantiSense and Netezza. It’s currently operating on a legacy platform of Oracle Discoverer for wholesale and supply chain operations, but intends to convert to MicroStrategy next year. PETCO has transformed from an older SQL server environment to an enterprise-wide data warehouse from Netezza and uses MicroStrategy reporting tools. Ascena, parent company of dressbarn, Maurices and Justice, is in the process of consolidating its data warehouse onto Netezza with MicroStrategy as its analytics tool.

Q: How do each of you actually use BI and analytics in your company?
Sherlock: It’s really been revolutionary for us; we’ve been able to take static paper reports into a much more self-service model. Some of the users are able to use drill-down capabilities and the power of the business intelligence platform. The person who runs business intelligence for me has a consulting background, and he refers to the four stages of an intelligent enterprise. I think first is visibility, second is information, third is insight and last [is] intelligence. He thinks we’re pretty much at stage three in that level of maturity. It’s a point where we are able to gain a lot of insights about our consumers and our products for our retail business.

Nell: At PETCO we’re sort of at a fork in the road. Reporting-wise … we’ve moved on to an enterprise data warehouse that allows us to report out quite effectively, but we can do much more in the area of dashboarding, exception reporting and so on. That’s still a challenge to us. On the analytics side, we have an amazing loyalty database that provides an inordinate amount of information about all of your [pets]. We know when their birthdays are, and we will sometimes call you and tell you when their birthdays are, if you forget — especially to come in and shop for something for them. It’s all about building a digital relationship and understanding the different elements of it, which is still a challenge to us.

Etzkorn: Using Janet’s stages as a guideline, businesswise we’re anywhere from stage one to stage three. Moving a group of apparel merchants away from their 4,500-page stacks of green-bar printout is a tough job, but we’re working on it, and we’re slowly but surely making progress. I’d say we’re at stage three in some of the operational areas of our business and … we’re creating some executive dashboards that really tell our business leaders by brand what’s working and what isn’t. With our store operations teams, we’re being very successful in rolling out dashboards to enable our field leaders to have information, not just data. As they walk into a store, they can see exactly what they should be asking that store leader about.

Q: What kinds of business impact are you experiencing as a result of BI?
Nell: The business impact to us is enormous. In the first place, as Karen mentioned, the ability to react quickly in the storefront itself is totally reliant on our being able to provide good, timely information. When you’re building an enterprise data warehouse, the real challenge is what we call the “slowest common denominator”: The slowest system providing data into the warehouse typically determines when you can get data back out of it. So it might be a limitation at times, but nonetheless we’ve been able to provide really good information to the stores and also to our supply chain operation.

On the analytics side, over the last three years we’ve been able to really start understanding the customers that shop in our stores. Those customers are not so easy to identify nowadays. Earlier we could say, “This is the profile of a customer who shops in stores, and this is the profile of a customer who shops online.” Today, the same customer shops in all our channels, but in different stages in the buying and decision-making process. That’s a critical element for us to understand, and we’re looking very hard at it.

Etzkorn: Each one of our brands operates extremely independently and autonomously and services the customer in a unique way. Our dressbarn brand primarily buys from the market, so our average purchase order volume is about 2,600 units. With 900 stores, the volume of purchase orders, and the volume of SKUs, is tremendous. Without an analytics engine we would never have been able to get down to the bottom and analyze what’s really happening in dressbarn by product, by product line, by store and so on. For maurices, the ability of field leadership to understand what they should be talking to the store managers about has enabled their productivity by store to go up.

Sherlock: When you’re using the more static reports … you have to look at things very hierarchically. The new system just gives managers a lot more insight. One thing I’ve noticed is that the business has become very reliant on having this data and, as Herman said, any chokepoint you have on getting data into the system so as to be able to generate reports is something you need to take care of immediately.

Q: You’re all kind of touching on this point, but let’s just dive into it. Business intelligence has plenty of challenges and plenty of opportunities, so tell us about some of each.

Etzkorn: There’s a big difference between looking at a report driven by business intelligence and just looking at data. Getting people away from looking at masses of raw data, and getting them to look at a dashboard with the understanding that they can drill down if they need to, can be a real challenge. Another challenge is the volume of data. I’m loading 86 million records weekly out of our Netezza platform into our JDA allocation platform. There’s no way I could do that without something with the power of the Netezza platform. I call it storage on steroids.

The opportunities are tremendous, from dashboards to the analytics to our ability to track and understand the consumer. At Justice we track our little girls’ birthdays so we make sure we send her a card so she’ll drag Mom and Dad into the store with her little 40-percent-off coupon and buy lots of product. So there are lots of opportunities.

Sherlock: You know, the promise of a data warehouse is that you’ll have just one version of the truth, but there’s really no such thing. When we import data into the CRM system, for example, it changes the calculation somewhat. I think the CRM system strips out employee sales, so if somebody looks at a report, they say, “That’s not the same thing that was in the data warehouse.”

Nell: We’re under pressure from the marketplace to understand our customer better, so we find ourselves being asked — and asking ourselves — to look at different kinds of data and look at it in new ways. A lot of this is driven by the channels; our customers tend to have done a lot of research on the nutritional needs of their pets, and they ask a lot of questions about the products. It can be very intimidating for our sales associates to be dealing with a consumer who has already made himself or herself an expert in the pet’s needs and requirements. When they come into the store it’s no longer a matter of getting them to buy from you, but of getting them to buy the right item. Making that kind of data available, both to the customer and to the associate, is a real challenge.

Q: Have you reached a point where business intelligence and analytics are channel agnostic — and if not, how long is going to take to get there? In sync with that, what are you doing about the enormous amount of unstructured data coming at you from social media?
Etzkorn: When we brought direct-to-consumer online, as it were, we were instantly integrated with our bricks-and-mortar business, so we’ve always been channel agnostic, at least from an information perspective. Social media is something that we’re quite frankly behind on, so I think that understanding how the customer operates in that space vis-à-vis our brand and bringing that data into the analytics is something we’re just barely getting started on.

Nell: I think we’re channel agnostic at this point with information, but there is still the challenge of mobile. In terms of that unstructured data, we’ve started down the road of what an earlier presenter [Kayak.com’s Terry Jones] called the “trust factor.” We have a Facebook page and various social media sites, and we’re in a step-by-step process of sticking our neck out there without trying to control everything.

Sherlock: I think it’s important when you’re looking at intelligence not just to be channel agnostic, but to be able to analyze the channels as they need to be analyzed.