by: John Caddell
In imitation of a Qwest ad from that time, I said, "I want it to run on any operating system, on any database, from any provider." I then glanced at him to make sure he wasn't winding up to smack me in the head. "You asked."
What I was trying to get across is that fighting the platform battle with customers is a certain loser. If they are a Unix shop, and you are trying to sell them Linux, or OS400, it's a nearly impossible task. It's better, frankly, to cut your losses and find another prospect for which your product is a good architectural fit.
Why is that so? A new book helps sort it out. "IT Risk," by George Westerman of MIT's Sloan School and Richard Hunter of the Gartner Group, discusses how companies can and should manage risk within their information-technology infrastructures.
And one of Westerman and Hunter's key points is that a company's foundation architecture must be simple and standardized. Such an architecture can be more easily protected from disaster, can adapt more quickly to changes in the business, and can limit data access to authorized parties more easily than a hodgepodge of separate systems, platforms and applications.
Like it or not, when you are trying to sell a nonconforming software product into a company that has built a simple, standardized IT foundation, you are trying to force them to accept a hodgepodge. And they won't do it.
Product managers need to manage the lifecycle of their architectures as well as the lifecycle of their functionality. It can be painful and expensive, but not as expensive as a good product that loses its market due to an outdated architecture.