Do You Need a Data Dictator?
Many companies are suddenly dealing with petabytes of information instead of terabytes. Keeping track of all that data and creating value from it, says Jeanne Ross, director of the MIT Sloan Center for Information Systems Research, may require more than technology — it may just require a data dictator.
Open a newspaper or magazine, click on a news story or case study and there it is, writ large: The promise of big data. What has brought about this sudden interest in data analytics?
Jeanne Ross, the director and principal research scientist at the MIT Sloan Center for Information Systems Research (CISR), says it is a confluence of events: companies are getting better at the processes that create their data, new technologies are creating new ways to access and analyze that data and business leaders are being bombarded with stories of others’ success with analytics.
Ross has been at MIT Sloan School of Management for the past 19 years. For the last four years she’s been the director of CISR, which focuses on how companies get value from information technology and on the digitization of processes and effective use of information. CISR has 85 corporate sponsors that fund the Center’s research, provide feedback and participate in events.
For her part, Ross examines the organizational and performance implications related to enterprise architecture, IT governance, business processes, disruptive technologies and business agility in an increasingly global, information-intensive world. She also leads executive education courses on IT management, working with such corporations as PepsiCo, McKinsey, General Electric, TRW, Cisco, Commonwealth Bank of Australia, IBM and Credit Suisse.
In a conversation with David Kiron, executive editor of MIT Sloan Management Review Innovation Hubs, and Renee Boucher Ferguson, a contributing editor at MIT Sloan Management Review, Ross discussed the issues facing companies today as they explore the potential of data analytics for their organizations.
How do you find the value in your data?
First of all, you have to know what is going to make you great. If you want to run yourself as a company that is data savvy, information savvy, analytics savvy, you need great data about your business.
That means somebody is dictating. There is somebody who says, “This is how we will define sales, this is how we will define returns, this is when we will register revenue, and we are all living by this rule. Until we do that, we don’t have data that’s useful for most kinds of analytics. We can still go out and buy demographic data and probably learn something quite useful. But if we want to know how to avoid stock-outs in our stores or what products are of greatest interest to a particular customer segment, we’re going to need the data cleaned up.” And that’s a major commitment. A lot of companies can’t get that done.
There are a lot of companies right now that are saying, “Why don’t we do something cool with analytics?” when they have sloppy business processes and equally sloppy data. And you know, they’re just never going to get there.
What is an example of a company that has the right approach to data?
Aetna is a good example. In 2002, Ron Williams is president, and he says, “Okay, we lost about $270 million last year. Let’s figure out what went wrong.” He brings in all of his senior execs, and he says, “Tell me about your part of the business.” And he said that every single line of business showed data showing they were making a profit. So, here’s the dilemma: Everybody says their data is showing that they made a profit, but the company is losing a ton of money. Something’s wrong.
Ron said, “The first thing I figured out is that I was going to be the single source of truth. I was going to dictate every piece of data, and you are going to use my definition of data.” What he got from that was totally consistent reports, and then he knew how to guide people through the process.
This wasn’t to beat up on people. This was to get the truth about the business so he knew what to fix. Because after his first set of meetings, he had no idea what was broken. It didn’t look like anything was broken. This is what I think the great leaders get right: they dictate a single source of truth.
That’s such a great example. But in that example, there is a crisis. In lieu of a crisis, what can help to get leaders to recognize that there needs to be a shift in the way they value data?
My best cases are all about a moment of truth where a company just said well, we are going to change or we are going to go under. A lot of them do create false crises. We’ve worked a lot with Commonwealth Bank of Australia. They didn’t have a real crisis. In fact, they were, by some measures, the largest bank in Australia. But management defined a crisis, which was to say, “Our operating costs are high and our customers’ satisfaction ratings are low. What this means is greatness is not sustainable.”
So they set two goals, one for operating costs and one for customer satisfaction. Then, they looked at structurally what was broken. And they started fixing the organization — the processes, the technology and the data — in a phased approach where they said, “We’re just going to keep getting better at certain things.” And sure enough, five years later, they are looking like an incredible powerhouse.
You do have to have a burning platform. You have to have this message that you’re either under siege by competitors or the world is changing and if you don’t get on this train, it’s just going to leave you at the station.
That calls to mind two potentially conflicting ideas. One is the idea that it’s really hard to get to be number one, but even harder to stay at number one. The other is: if it ain’t broke, don’t fix it. How do you create a burning platform, when there is a tension between those different mindsets?
That’s a great contrast. Exactly right. How do you do that?
P&G was very much in this position in 2000. Results were good but trends, particularly cost trends, were bad. They had grown to 4 billion consumers. But the growth markets were in developing countries. Those markets really challenged their business model. In Turkey, consumers often want to buy one diaper at a time. In India, many consumers buy little, single-use shampoo bottles. And suddenly, P&G is saying, “Wow, we want to keep growing and our investors expect us to keep growing, but we are really hitting some limits to our standard processes.”
That’s their dilemma right now. Wall Street is saying, “Guys, you need to grow.” And yet, if you look at what they are capable of doing, you think, “Man, a lot of companies would kill to get there.”
It’s not clear how they get to the next level. What do you do about that if you’re P&G? That’s really hard. Except that you know two things. All the analysts are telling you that Colgate and Unilever are not having the problem you’re having. That’s a very interesting problem.
Could you explain the idea of sacred data that you talked about in your recent e-Chat with AllAnalytics.com?
First, I should tell you I learned this term from Tom Nealon, the former CIO at Southwest Airlines [now a board director there], and he got it from Charlie Feld, who was the CIO at Frito-Lay and a variety of other places. But the idea is simply that your data is not all equal, and if you treat it as if it is all equal, you’re never going to get anywhere. So figure out first of all, what’s the single most important thing.
Tom said that at J.C. Penney’s, which is where he went after he was at Southwest, the purchases were the single most important thing. You needed to know product, basically. At Southwest, it had been the customer reservation, because that runs through the entire system. At UPS, it’s the package data. By the way, UPS is brilliant at analytics.
There are other things that matter. But you have to let the others go until you’ve got the single most important data fixed. Some of these fixes will not be as elegant as you would like. What you’d love to do is just rip out all the legacy that touches the customer record itself and start from scratch, but you can only do that if you’re willing to turn it all off. Can you turn off all your systems for a year, rebuild it? You can’t.
We started learning this in the ’90s. I think we were surprised how hard it was. It was kind of like, “Yes, let me fix my platform. Quick, put in SAP.” And when that didn’t work, there was this depression on top of major financial catastrophes. It’s like, “What do we do now?” The inclination is to say, “Well, ripping out the legacy and replacing it with an ERP doesn’t work, so we’re not going to do that.” But if anybody’s answering with “Oh, well, then why don’t we do data analytics instead,” they are in for a shock because they’re really not going to have the data to do the analytics — and they won’t have the discipline to implement any new processes required to take advantage of the learning from analytics.
If your core data is bad, you can do analytics around the edges, but you’re never going to figure out how to avoid stock-outs or better serve customers. You’re not going to figure it out because you don’t have the data. There is just some data that, as messy as the process is, you’ve got to get right, and if you don’t give that data all the attention it needs, you’re not going to get there.
What do you think is responsible for the new wave of interest in data and analytics?
Companies are starting to get better and better at storing data and finding easy ways to get to it. So the technology has made things possible that weren’t possible before. And companies have learned how to use that technology. There is much more of a readiness than there was years ago, when we learned that Capital One was doing these incredible analytics that was making it possible for it to customize credit cards to very small demographic segments. Now it’s like, “Wait a minute, shouldn’t we all be doing that?”
So I think there has been an announcement that analytics has real potential at a moment when companies are saying, “Wow, we have data that we never used to have, and we have access to that data in ways we never dreamed of. And then, there’s all this external data that we can go out and get. So yeah, let’s analyze it. Let’s get smart. Let’s do something our competitors can’t do.”
It feels like the moment for analytics, though it’s still much more about the promise than about the reality. Some of the things we learn about our customers or products or more general demographic trends have the potential to stimulate some creative opportunities, but if a company doesn’t have a vision for how it will succeed, these isolated findings will tend to distract rather than drive business success. Businesses need core competencies and they need digitized platforms if they want to succeed in a digital economy and address the learning from data analytics. That’s the hard part of analytics. Anyone can hire a quant jock to analyze some data. Putting it into action requires enterprise capabilities.
When would you say this new wave of interest in data and analytics really began?
The real excitement has been the last couple of years. It’s the “big data” thing. It’s that suddenly we have petabytes. What’s interesting is that organizations go from terabytes to petabytes, and both of them are unimaginable, right? [A petabyte of data is a million times bigger than a gigabyte and a thousand times bigger than a terabyte.] And suddenly organizations find, “We’re in petabytes. And oh my goodness, we’d better get value from it.” So part of it is the hype. Part of it is that it’s really possible to enable our people to work smarter and to make strategic decisions.
The hard part for a manager right now is focus. We just don’t know how much or how big or what next turn we should take. And the sense is, “I shouldn’t just sit around and watch because other companies are getting smarter and I’m watching.” That feels like a bad formula.
Comment (1)
Sundarrajan LN