One of my newsletter readers sent me an interesting question:
“I was wondering if you have insights on the attitude of managers on the business side towards the re-alignment of IT. I can imagine that managers at the business side tend to resist giving up (overly) customized IT, since the short-term performance of their individual units is so dependent upon (overly) customized IT. Asking a manager to cooperate with de-alignment, when no alternative for ‘less IT’ is available yet, means asking a manager to sacrifice his/her short-term performance.”
This particular reader works in a large multi-national consumer products company, but I believe her concern is a common problem these days. Companies are cutting back everywhere they can, and it’s widely believed that centralizing IT can save money. Here are my thoughts on the subject:
1. Change can be Painful, But We Still Need to Change
There’s always a transition cost to any change, as people go through a learning curve of changing from one process to another. So in a sense, all managers have to constantly cope with “sacrificing” short-term performance in the interest of longer-term gain. That’s just the cost of doing business in a changing world.
2. Centralization Can Offer Advantages to a Decentralized Operation
The implication of the reader comment, however, is that a local manager who has his/her IT support changed as the result of an IT realignment gains nothing. The reader also refers to the change as “de-alignment,” when in fact it’s more of a realignment to meet higher-level business goals instead of optimizing local departmental goals at the expense of overall business objectives. If a realignment is done correctly then the local manager can actually benefit from the change.
Let me sketch out what I see as a common scenario; this may or may not match your own situation. In my scenario, IT is currently distributed across multiple organizations and probably multiple locations. It may be manufacturing plants or companies that have been acquired, or it might just be different departments in a single company and location. In any case a decision has been made to “centralize” IT. Such a decision is usually made as a move toward economy, since economy of scale is virtually the only reason ever given (or needed).
The local department manager sees control shifting from his/her local IT people to someone else in a central location. But what the manager doesn’t see is that:
- The local IT people will no longer have to waste time handling basic infrastructure needs that are better handled by a larger central staff. This really is an economy of scale, since it’s difficult for a small local staff to have the depth of expertise required to properly support a broad range of infrastructure products.
- Off-the-shelf systems like ERP that the local IT people couldn’t afford will now become possibilities. Those off-the-shelf solutions will probably do a much better job of handling local needs than the low-budget built-from-scratch custom solutions that previously existed.
- There will now be a better career path for IT people and so it may be easier to hire good IT people into positions where they’re needed.
- If systems are standardized across different departments and locations then it will be easier for local non-IT managers to move people from department to department or location to location without extensive retraining.
- With standard systems it will be much easier for people in different departments and locations to share approaches and process improvements. And system enhancements can be used by everyone – not just those in the location doing the improvement.
The problem with distributed IT is that the solutions built locally are so highly customized to suit the needs of the individual people in the local departments that they aren’t ever any better than the ideas of those individual people in the local departments. Some of those individuals are brilliant, and their IT shows aspects of this brilliance. But most localized customization just reflects the preferences of average individuals, and there’s no big advantage in that customization. Centralization offers the possibility of bigger thinking, bigger solutions, and – hopefully – better solutions.
3. Transitions should be Managed Carefully
If the transition from decentralized IT to centralized IT is done correctly, then there won’t be a sharp decline in performance; there will just be a gradual shift from using overly customized solutions to using more generalized but better solutions. In an ideal transition, there will be an inventory taken of the local solutions that may be replaced, and the best ideas from each local area will be added to the overall corporate solution.
The biggest issue with IT centralization is a perceived lack of responsiveness and personalization. Local IT usually provides a lot more “hand-holding,” and the locals get to know the people doing the support. Losing the hand-holding can be traumatic, but even this trauma can be minimized if local IT support is continued at some level. Centralizing IT doesn’t usually mean that all of the local IT people move elsewhere, and a good IT organization will encourage local IT people to maintain the relationships that they have created, even if less time is spent on doing support.
4. Any Centralization will Likely be Followed by a Future Decentralization
The business needs to seek an equilibrium between a centralized and decentralized environment, and if the equilibrium isn’t achieved, then there will be a backlash and the cycle will reverse. I talked about this in my June, 2005 newsletter article, and human nature hasn’t changed since that time.
But don’t look at this centralization/decentralization cycle as just an IT problem. It happens in everything we do, from government (federal control versus local control) to religion (central control versus local control) to family (parental control). We all want the benefits of a centralized organization without giving up the freedoms we have as individuals or small groups. It’s one of life’s continuing struggles, and perfection is not likely to be achieved.
IT centralization works best when each aspect of IT is considered separately. Centralize those things that will benefit from centralization, typically things like data centers, help desks, purchasing, standards and maybe development and support of corporate-wide software and systems. But for those things that show less benefit from centralization, you should seriously consider leaving things the way they are, or maybe just putting in place a way to better coordinate decentralized IT people rather than replacing them with a totally centralized alternative. Centralization offers economies of scale, but not everything in IT can benefit from those economies.
Related Postings and Articles: