When I was a kid, I played with mercury I harvested from thermometers. (I also made lead weight belts from scraps of lead I retrieved from gutter repairs we performed on my grandmother's three-decker flats in the factory town of Brockton, MA where I grew up.) My wife wonders if my parents had a death wish for me... The vivid thing about playing with mercury (not that I recommend this toxic pastime) is the way it squirts away from your finger into smaller and smaller balls as you try to retrieve it. Each blob explodes into hundreds of pellets which fragment into thousands of bb-sized silver balls, and on, and on.
Information in organizations can be similar. As soon as you let information enter your organization -- it is copied, broken into pieces, slightly changed, moved around, and it expands into more and more fragmented items like a giant vat of mercury squirting into every nook and cranny of your firm. I think it was Aaron Widalsky, who wrote a paper called something like, "Information as an Organizational Problem" in which he argued that the primary purpose of organizations is to rid themselves of information by turning data into information and then into action. He reasoned that if information floated around the organization -- unprocessed -- it would take more and more information worker's time and information workers themselves, to handle it. I submit that many organizations are chest deep in information fragments and organizational actions may take on the speed of a medieval knight swimming in chainmail -- due to all the information fragmentation and its derivatives.
The usual executive solution for the concomitant information overload that comes with the fragmentation of the information flow in an organization is to force "one platform" for all critical information in the organization. This centralization is often useful, especially when there are many redundant systems and processes, and once an organization gets to one version of the truth -- it can be very powerful. W.W. Grainger, the $5 billion distributor runs all of its stores off of one core platform, which allows them to provide end to end service across every one of their 500 plus branches across North America. You can be in Watertown, Massachusetts at the Grainger branch near my home, and not only find out if you can get a 2.5 HP GE electric motor in downtown Chicago, but you can reserve the actual inventory at that moment -- because they have one system for information and for inventory, across all locations. Dick Keyser, the CEO, believes that his single platform enables his firm to deliver better variety and service, and handle more complexity than his competition. His single platform allows Grainger to handle a higher level of complexity than the competition at a lower cost level. It is vital to keep in mind that Grainger's journey has been long and focused. Keeping a core platform in place has been an exercise in constant reinvestment, and perseverance. Keeping the information platform in line is not like cosmetic surgery, where you can enter the hospital on Tuesday and walk out New and Improved on Wednesday, it is more like orthopedic surgery, in which you break a bone, reset it, break it again, reset it again. It is a constant management focus -- in order to get it done right.
Firms that have not had the ongoing discipline and investment often try to catch up by slamming in a centralized information handling system -- such as SAP, or other Enterprise Requirements Planning systems that promise one version of the truth. The salespeople who sell these systems sell the hope of "control" to an executive thirsty for more visibility and control of their complex organization. This thirst comes from the fact that the essence of being an executive is to be held accountable for those things which you cannot control in detail. Thus, many executives feel the angst of being held responsible for actions they cannot see in detail. This powerful lever is why Enterprise Requirements Planning salespeople are successful at selling things that cost hundreds of millions of dollars to buy, and much more to implement. Clarity is a seductive promise. Yet, still today, the majority of projects designed to centralize and standardize core information in an organization often (60% or more of the time) fail.
Some of these failures are due to incompetence, but most, I think are due to the fact that companies are trying to control information at the center, when they should be thinking about thow to push both the information and the decision rights, to the skin. When an organization can allocate its decision rights to the edge of the organization -- then all starts of interesting things start to happen. In their landmark book, The Service Profit Chain, Heskett, Sasser & Schlesinger showed that by giving front line service workers more autonomy (e.g. decision rights) and more incentives (e.g. a reason to make the right decision) it was possible to make the firm grow faster and more profitably. The fundamental source of the economic value came from the fact that the cost of the layers of oversight not only was more expensive than it needed to be, but if you took that same money and gave it to front line workers to create value, the net impact was that the business grew and grew more profitably. Their examples came largely from organizations where the unit of production was near the edge of the organization -- like franchise businesses (Len Schlesinger took these lessons to Au Bon Pain where he was COO for a many years before going to The Limited), and hotels, which even though they often have centralized reservations and brands, the local property is where the service is "produced".
From an information standpoint, the "better organizations" allowed the employees to process the information and its resolution into action, at the "skin" of the organization. Those closest to the customer must deal with the information and its implications. By allocating more and more decision rights to the skin, less information floats around, and more gets done. From this point of view, all of Customer Relationship Management systems (CRM) is simply used to move the relevant context for the data to the decision site. Why do customer service representatives at Harrah's know my theoretical profitability? Because you want to make sure that the right information is at the skin of the organization so that they can do the right thing.
Many executive efforts, from the hulking bureaucracy of Homeland Security, to the implementation of many ERP systems are simply an attempt to "simplify" the information problem by centralizing it. Yet centralizing information often makes it more complex, because information is like mercury, and will constantly change, morph, and multiply. The more effective action it to simplify as little as possible, and allocate the very best information, and information analysis tools to the edges of the organization -- toward the skin -- where incoming information can be warded off by decision making, with the same clarity that my white blood cells kill off pathogens -- without checking in with my brain first.
More seductively, the promise that information will be simpler if it is centralized seems to makes sense -- for we think of information as if it were a thing -- like cheese. If you put all the cheese in one fridge, it is easy to slice what you want when you want it. But information, especially customer-driven information is organic and lives among and between you and your customers. In such an organic situation, it makes much more sense to provide the very best analysis tools, and most information access to the very skin of the organization -- and only centralize a very few, and very important set of decisions. Otherwise, the very act of centralization will create so many information back channels, and reinterpretations of the situation, that decision making will be slowed even more.
So, what's a poor boy to do? First, make sure that your information strategy matches the information reality in your business. When Bob Nardelli came to Home Depot, he centralized the decision making on merchandizing and buying. He believes that the efficiencies of centralization will help him win the day. At the same time, he disempowered the store managers. I believe that this is a fundamentally flawed strategy because the efficiencies of central buying and purchasing will be eaten up entirely by the increase in information and coordination costs, and the concomitant loss of motivation by the local store managers. Unlike Grainger, Home Depot does not need total visibility to meet the needs of its customers. Meanwhile, Nardelli has been given over $200,000,000 in compensation, while the stock has gone nowhere! I expect Home Depot's slide to continue until they change their information and decision making approach.
Second, think through what can be codified (like CRM, or repair manuals, or refund protocols) and give the very edges of your organization more and more information and autonomy. Even think about how you can empower the customer. GM has just begun to offer to OnStar customers that they can get an email about their car's performance. They have had over an 70% opt in by customers -- which is a rate of adoption an order of magnitude greater than even the best marketing campaigns. If you put the information at the skin of the organization -- even into the customer's hands -- you will win.
Third, think through any centralization activity carefully, and for every increase in centralization -- create a concomitant decrease in centralization. In an information world, it is tempting to centralize more and more, because it is technically possible. However, resist this natural urge, for when things are overly centralized, the organization will reroute around the control. Trust is cheaper than oversight -- use it whenever possible. It is also technically possible to decentralize more than ever before because customer status, problem solving heuristics, and service recovery options are more easily shared with today's information systems than ever before. Give out the power whever possible. If you over-centralize instead, you may be left with a voluminous description of the situation with very little understanding of the reality on the ground. When that happens, you will be able to empathize with Michael D. Brown, the head of FEMA, during Hurricane Katrina -- who knew less about status than did CNN. His organization simply cut him out of the loop.