How much freedom should business-unit managers have in deciding which information systems to buy? It goes against my anti-bureaucratic grain, but I’m afraid my answer to that question is: Not much. The benefits of consolidating and standardizing a company’s IT assets have grown so great – and the penalties of incompatible systems so onerous – that most companies should be moving toward a centralized governance structure if they don’t already have one.
Nevertheless, the central vs. decentral question is always a tough one to wrestle with. A good overview of the relative merits of different governance models is provided by Peter Weill, director of MIT’s Center for Information Systems Research, in this brief research paper. It reports on the results of a study of 300 companies, which found that the IT governance structures of top-performing firms vary widely depending on their corporate goals. Companies focused on profits exhibited greater centralization, those focused on innovation and sales growth were more decentralized, and those seeking to maximize asset utilization had hybrid models. The findings are summed up in the following chart from the report:
The choice of a governance structure has, as the chart implies, always entailed painful tradeoffs. You couldn’t minimize costs, maximize asset utilization and spur innovation all at the same time. But I think those tradeoffs are becoming less severe. As IT infrastructure becomes more flexible, business units will increasingly be able to cede control over IT assets without sacrificing their freedom in applying those assets to their particular business challenges and processes. Standardization no longer means a loss of flexibility.
It’s hard to argue against cost saving on IT centralization. Yet one of favorite CTO’s had this saying which I believe. “Managed diversity has great value in the enterprise.” His organization which was full of scientists and researchers benefitted greatly from letting those folks pick the right tool for their job and the one which they enjoyed using the most. The increased productivity was well worth the cost of multiple platforms. Since I was Director of Federal Sales for Apple, it certainly made my team’s job easier. It was just this decentralization in organizations searching for innovation that allowed Apple to keep a toe hold in federal markets when most enterprises were throwing Apple out.
The one thing which argues against central control and a single platform is the real security advantage of having diverse platforms. While it is hard to quantify the value since few exploits have gone beyond clogging mail servers, many of us believe that it is just a matter of time. Then perhaps the long term view that diversity equals real security might have a shot. Having a different system running another code base which continues to work under a serious attack might look inexpensive some day.
Because of the importance of my internet connectivity, I have both cable modem and DSL. Today a backhoe cut the cable modem line. The backup system looks pretty good today.
The centralized vs decentralized is becoming a bit boring and tends not to push the real debate that needs to happen with IT organizations. The real debate centers around the type of people, leadership styles, innovation capacity, ability to collaborate and other difficult areas for IT to grapple. The discussion is also boring, because most IT senior executives and CxOs in many cases are using references from “past experience” which can be helpful, but only as a context–not what should actually. For example, I am working with a client whose five business units scream and clamor for autonomy and thus would love to have a purely decentralized model. Yet, while the units perform the same operations, these units do not collaborate with each other, they do not share information in an easy to use way if at all. By contrast, due to the low pay for this client’s industry, there is no centralized leadership in place to guide a more centralized model. So what is the answer–continuous demands to outsource because of soaring costs; of course this merely treats the symptoms and not the problems.
In sum, the centralized vs decentralized debate tends to not take into account more insight discussions about the type of behaviors the organization wants to promote or the competencies of the people in place. We’ve all seen the story when the new CEO comes in from a centralized model to lead an organization working just fine within a decentralized model. The response to the media from the CEO is, “we will take advantage of economies of scale” and “release new value” for our shareholders. When the plan does not go as far as she thought it would, a few heads roll then a year or so later, the CEO gets axed. The debate is often predictable and requires more in-depth analysis by consultants and managers to truly understand what is best to capitalize on the strengths and talents of the organization and its people.
IT governance and identity management
As the new MIT survey (via Nicholas Carr) shows, good governance does not happen by an accident. This is pretty topical, as I am more and more convinced that it is actually a company governance and operational model that drive…
Study on IT Governance Structures
Wayne Hall and Nicholas Carr write about a new research brief from MIT that covers various IT governance structures. I wish that I could write a lot more about this right now because I absolutely love this stuff, but for now I figure I should at least …
IT governance and effciency
Somehow or other, I ended up reading this article on Nicholas Car’s blog about governance of IT purchasing in organizations. It highlights an important point: the governance structure should be focused on the objectives of the business, not as a