No Small Software Purchases

We seem to have entered a new age of Large Software Purchases.

When I started in this business, the mainframe era was ending and the super mini computer (yes, our terminology was lame even then) was on the rise. Departmental servers, in the shadow of the might core computer, sprang up everywhere. The hardware was expensive ($20K-$50K) and the software generally quite expensive too.

The computers started becoming personal and the range of prices for software used by businesses grew much, much wider than it had been before. There was even overlap between what people bought for home use and what was in use at the office, which surprised all of us, even Microsoft.

Along came Linux and Open Source and the price landscape changed again, for those daring enough or strapped for cash enough to take the plunge.

But recently I see a strong swing back to centralized purchases, central servers and corporate software. I see executives asserting themselves at the senior level, making large software purchases from their 50,000 foot perspective.

This shift is not inherently bad, but this central decision-making is often accompanied by a very unfortunate policy of "no support for local requirements." HQ buys it, you use it. HQ did not want to get mired in the details, so they do not know what, if any, local requirements there are. They even hint that local requirements are a function of local management incompetence.

I have seen a proliferation of large systems which are not well suited to the task at hand, or systems which are not properly configured for local needs, or systems which are both.

This trend has been explained to me as an attempt to conserve management time, which is often a company's most expensive resource. (I find that fact disturbing and there is probably a rant in there somewhere.) It is more efficient for there to be fewer, larger IT decisions and it makes sense to have those few, large, decisions to be made by the decision-making experts, without an undue focus on minutiae.

I can understand that the sort of process problem that I make my living solving just isn't important enough to warrant senior executive attention. But I think that something has gone wrong when large swatches of the organization are making do with inappropriate or misconfigured "enterprise solutions."

I can grasp that sometimes a local problem should be left unsolved for the global good. But I think that such calculations should be explicit and revisited; instead, I see local issues being ignored in such large numbers that it is hard to believe that the sum total of all that procedural losing does not outweigh the cost of acting to prevent at least some of it.

I realize that economy of scale does come into play sometimes and that having a moderately sized central department can be more effective for some purposes than myriad local departments. But I am suspicious of this model when so often there are local needs dictated by regulation, competition or regional variation in the customer base.

I can see that large companies are complex and require some metaphoric distance before you can see the whole picture. But I suspect that this very complexity argues against the "one size fits all" philosophy. This feels to me like the old trap of trying to solve personnel problems or process problems with software. That does not work and I suspect that we will find that large software systems imposed from afar won't work either.