Saturday, February 13, 2010

Enterprise Software Shouldn’t Be So Complicated

In an era when most kinds of computer software are getting simpler and cheaper, one category remains stubbornly messy. That’s the software that big business runs on. The old joke is that it’s called “enterprise” software because it costs as much, and breaks down as often, as the U.S.S. Enterprise, the starship from Star Trek.

As a sign that the state of this category is perhaps a little worse than people say, it is getting harder to sign up growing businesses for enterprise software installations. Small business that grow to the level of about 50 to 100 employees, the point at which it’s possible to make a case for replacing the disorganized repository of electronic documents with some kind of enterprise software approach, aren’t considering enterprise software at all, but are tending instead to pay for a cloud computing approach.

Cloud computing is the most complicated computing architecture ever envisioned, and it tends to be priced accordingly. One of the most prominent examples of cloud computing is the Sidekick service provided to T-Mobile users by Microsoft. The service provides access to a tiny subset of the kind of software that comes free on a $450 desktop computer, but it costs much more — it’s part of a subscription for which users pay that amount every three to six months. Even at that price, the Sidekick service failed spectacularly five months ago, remaining partially offline for about two weeks and losing practically all the data that users had trusted to it.

You would think that kind of track record would make businesses hesitate, but I am afraid the track record of enterprise software is worse. Everywhere you turn, there are stories of enterprise software installations that went two years longer than planned and $10 million over budget. It’s enough to keep a business executive awake at night.

Enterprise software shouldn’t be so complicated. There’s a huge market out there for anyone who can persuade a medium-sized to large business to adopt the same kind of simple, stable view of business data that small businesses and investors use. That kind of enterprise software would have only a tenth of the complexities and idiosyncrasies that we’ve come to expect from enterprise software — and it could come at a tenth of the price. It wouldn’t have to cost a million dollars to develop or install, and it wouldn’t have to take weeks for the employees to learn. And most importantly, big businesses wouldn’t be misled by the view of their businesses they would get from this software, the way they are with the twisted, distorted picture that current enterprise software provides of their businesses.

Think of enterprise software as a way to look in the mirror. Big businesses are saying, “Our business isn’t flat, so a flat mirror won’t do. We need a big custom-molded mirror, no matter how much it costs.” The more savvy growing businesses are saying, “I don’t need a mirror, because I’ve got this webcam hooked up to a server in India that sends me snapshots by e-mail.” Meanwhile, everyone else in the world has this problem solved — they use any flat mirror that gives them a reasonably accurate reflection. It’s a solution that large businesses could adopt too, if someone could persuade them that it would work.