Ann All spoke with Ron Swift, vice president of Cross Industry Solutions for Teradata, global leader in data warehousing and analytic technologies.
All: Do companies sometimes have trouble identifying the specific business needs they want to address with their data warehouse strategies?
Swift: I think there are three spectrums or views. The first one, a company that is in trouble or has to change or is in pain, generally has a project mentality or a specific area mentality. The executive is sometimes forced into making a decision. A lot of times when (companies) are in that position, they tend to want to do it in the most economical fashion, not necessarily the correct fashion. They take a short-term approach, putting a tourniquet to stop the bleeding on the leg instead of doing heart surgery. There are some very finite things they have to do in a short period of time, perhaps two budget cycles at most.
The second kind of company is at the other extreme. Everything is going really well, their competition is not beating them, they have a pretty fair infrastructure for information and decision making, but they want to introduce changes in terms of taking the company - to use a trite term - to the next level. They tend to want to have some kind of a strategy where they are going to make gradual changes across the business.
And then there is everybody in between. So you almost wind up with 20 percent at either end of the extremes and 60 percent in the middle. The center group tends to say: "We understand we've got these five, 10, 50 problems. We're going to solve them one at a time, and we're going to have different people working on them, and we're going to try to work with various unit or area executives so there is some coordinated responsibility or accountability for resources expended, or the change that occurs and the success of that change." The biggest problem with this approach is, they change the people. What you tend to see is a lot of change management not just in the process, but they tend to change the people as well.
All: Are companies looking to leverage their data warehouses to improve customer service?
Swift: Some companies have an enormous problem, relatively low loyalty percentages, because of product or price not because of quality of service or experience. The better companies on the far end of the extreme are working to optimize the customer experience. We cooperatively invented the term: CEO (customer experience optimization). So it's not just about selling (customers) stuff, it's about managing the relationship. Those companies don't struggle with change because their customers are staying for reasons other than just product or price.
There is a percentage that is struggling to figure out how to provide good service, how to be on time, on schedule, within budget, all those normal things, and at the same time, how to keep the customer loyal. Many organizations, including CFOs and CEOs, don't really understand the cost of replacing customers. That requires the advertising costs you've already been incurring to brand yourself to the customer, getting the lead on the customer or qualifying the customer, and converting those initial interactions to an actual order. And usually it's not the first order that matters, but the second, third or fourth one that is going to turn into regular revenue. Unless a company is totally hacked off at a supplier, it is not going to give you the full order. It'll let you participate as a competitor and try to migrate to you if you perform better. If you can't perform better, deliver on time, be friendly, all of that, you aren't going to get the order. So this loss of a customer is gigantic.
Then once you get the customer, how do you keep him? A lot of people think if they have a new product or a new version, they'll keep customers. That's not how you keep customers. You keep them with good experiences, quality processes and quality deliverables, whether it's product or service. At the same time, you have a cohesive organization that is working in synchronization. Unsynchronized organizations frustrate the heck out of customers. A bank that doesn't know you have a million dollars in wealth or doesn't know your volatility on your credit card or doesn't know what you are doing in the branch knows nothing. People think that CRM is selling products, tools, toys. But that's not CRM - it's CPM, consumer product marketing.
All: What are some real-world examples of customer experience optimization?
Swift: Have you ever been to Best Buy? They keep track of what you buy and send you reward points in the mail within the week. They mail you stuff right after you shopped to keep you coming back with offers that have an affinity for the products you just bought or maybe even the trends of products that other people have bought over the last two or three years where you haven't jumped on a trend. Macy's sends bulletins to their cardholders saying we want you to come in because we have an offer especially for you.
All: So you're talking about targeted marketing?
Swift: Targeting has traditionally been very large customer segments, exactly the same product, exactly the same price, all at the same time. Personalization is using the data in your system to personalize the offer, price the offer, package the offer and then be relevant even in how you message the offer. It's very different from targeting. That's database marketing from 20 years ago vs. what some companies like Best Buy are doing today. If I send you a promotion with a picture of me (an older man), you probably wouldn't buy it; but you might buy it if I sent you a picture of a woman with your characteristics. That single change can result in a 5 percent to 10 percent difference in sales. Multiply that 5 percent or 10 percent a million times; that's big money.
All: Just getting more data about customers doesn't seem to be the answer. Do we need better analysis tools or better interfaces for interpreting the data?
Swift: What are we going to do to help build infrastructure? These companies will have more customers, more data, more processes, more products, more relationships that have to interrelate with each other. So these systems for executives are not just these initial systems to port off these reports. A lot of people do nothing with reports. They scan through, look at the bottom line, look at a color graph, see some anomaly or trending - but they've got a wealth of information behind it that could be put not just in a different format but could be cohesively integrated in a different way.
There are a lot of studies out now about dashboards, and I think they are totally missing the point. The real issue is: If I have to look at a report and think about it, then I've wasted a resource to produce it. A report should show you instantly what you're supposed to do. That's the use of the data warehouse. The report shouldn't show you the problem, it should show you the solution. Red and green doesn't help me. What some people do in this business intelligence world is, they say, "We gave you the report, now you drill down and find the answer." That's a little ridiculous for someone who makes $300,000 a year, $800,000 a year, $2 million a year, to sit at a PC and try to figure out what the problem is. But they were taught that way.
Visualization offers a wonderful opportunity not just for the executives, but for the people who present reports to them, to solve problems that would never have occurred. What's striking is that CEOs and CIOs don't know how to read these charts. It's called a dial chart or a circular chart, and it shows the company's sales performance for the whole year. They can do it by product, by customer set, by store, whatever. I can see a very finite community of customers. (Visualization) is 10 years old. We've been doing it since '97, '98. The software was very specialized at first, and MBAs weren't taught how to read these kinds of charts until just recently.
All: Why aren't more companies doing customer experience optimization?
Swift: They get stuck after the analysis and prediction stage. There are two more stages beyond this. First, if you can take a look at how people set up requirements for business intelligence, you are going to come up with another set of requirements beyond just generating the report. If I know how to analyze this and what these reports mean to me, and I've got the data, I should be able to figure out some commonality of my customers and what they will do next. Therefore I can do the prediction. To get them over that gap is not that big of a jump. It's not about the technology; it's about them believing the prediction.
The example I use is if a company gets 85 percent of its predictions correct, you need to ask a serious question of the CXO: What happened to the other 15 percent? When somebody tells me they have 78 percent customer reduction, I ask what happened with the other 22 percent? Couldn't we be at 92 percent? When you get to that kind of thinking, it's not about building a statistic, it's about resource allocation. They get very good at predicting what will happen in the future, and they will allocate their resources very differently. I operationalize that prediction by putting it into the flow of daily work.
This is not a new idea. If you go back 15, 17 years in the retail industry, there was a company in Bentonville, Ark., that built a system that told it how much product it needed to buy, and where did it sell, and what was on the shelf and in the warehouse. They got this direct information to 800 suppliers and said, "You manage what's on the shelves. We don't care how much toothpaste we have. You guys should know that. And that will affect your production and transportation costs to get your products to our distribution centers. So you can cut your costs, and we can get the products from you just in time." You use the flow outside the company to benefit the company and its customers.
All: You stress the importance of doing a business impact study. Why is this so important?
Swift: If you cannot explain to an executive what the capital impact of what you want to do is going to be, then you're dead. We do a lot of these. We don't do a cost-benefit analysis; we're into looking at what the impact to the business will be in a number of different areas. I buy this hardware and save the company money. That doesn't help. I want to reallocate resources into growing top-line revenue. If you want to do it right, you figure out how to invest in technology to drive 2x, 5x, 10x growth.