Paul Murphy has been posting excerpts from his book in process, "The Board Member's IT Brief," regularly on his blog, and today's entry covers a subject near and dear to our hearts here: Hiring IT Management.

Paul's primary directive in this matter is to "pick someone whose skills align with the technology you have or want." His reasoning is that technical expertise is rare and the candidate can be educated about the company's market vertical by other executives. The danger in not doing this, which he illustrates with a tale of woe from a client, is that the non-aligned CIO may do some technically stupid things and no one else in the organization will have sufficient IT knowledge to catch and prevent this.

You've probably already spotted the major flaw in this logic: what are all these technical incompetents doing picking the technology they "have or want" in the first place? They have all the knowledge and skill in the first place to evaluate and select the correct application and system architecture, but they need to be careful not to hire someone who will pull the wool over their eyes and do a poor implementation job with it? Isn't picking the technology something that you are supposed to hire a CIO for?

Paul's example is of a hospital he consulted with which had selected a Unix, smart-terminal based information system which was installed to rave reviews. Then the organization brought in a CIO with a System 390 background, who didn't understand Unix, and promptly ran the IT department into the ground by staffing and running it as if he were still in a mainframe environment.

Clearly this was a poor hiring decision, but it wasn't poor because the CIO was from a mainframe background. It was because he couldn't adapt; and that could have been just as true if they hired a Unix expert. The time of reckoning may have been pushed back, but nothing is as sure as change in this business, and would have come the day that Unix was the wrong skillset as well... if you're looking at things at that level. But I think that's the mistake in this approach. I think he's talking about a job for a CTO, not a CIO. After all, how many enterprises do you know today which only run Unix, or only run Windows, or only run a mainframe? What was the hospital to do for a CIO if it was instead a healthcare system with an umbrella over a dozen other operations, each with specialized needs and systems? Hire a dozen co-CIOs with the requisite expertise?

I think that Paul has it entirely wrong, and that picking a CIO on the basis of the technology they are expected to work with is the kiss of death. Oddly, our reasoning is similar, but Paul doesn't extend his logic over time. The primary objection that he lays against picking someone unfamiliar with the existing technology is that they will instead gravitate to another technology with which they are familiar, and hose up the existing installation along the way. That's a risk, no doubt... but it's also an argument against hiring someone who specializes in the existing technology. Technology, and business, evolve. Paul correctly cautions that the CIO's unique technical expertise make it difficult for other executives to evaluate his or her technical proposals. Certainly that factor can allow a System 360 guy to make a hash of a Unix shop. But it also can allow a Unix shop to perpetuate itself even if something better comes along.

The CIO should be the resource for picking the technology you need in the first place, and that means hiring one who doesn't specialize in anything, but instead understands your business, and business processes, and can judge technologies impartially. He or she needs an open mind and the ability to hire the sort of specialists that are necessary for a smooth system implementation and operations. The CIO is the bridge, the person who the board and senior executives can trust to provide the sort of technical oversight that Paul, eventually, asserts they are incapable of. You have to have technical expertise, but if you think that is allyou need (or even the most important thing you need) you're walking into a blind alley. Find a CIO you can trust to steer you toward any technology you might need, not just the one that they happen to know.