Technologys Barriers to Exit

Opinion: When purchasing technology, what happens beyond the costs of acquisition and operation?

Most businesses have become pretty savvy about the true cost of technology—though not nearly as savvy as they should be.

Three things are critical; they have two covered.

The first thing—cost of acquisition, aka barrier to entry—was obvious from the get-go, and every company I know negotiates like crazy to get the best price. So far, so good.

Next up: the ongoing cost of running everything. They have this one down, too. How much will it cost me to fire up all these microprocessors? How much to cool them all down? How much floor space will I need? Whats tech support going to cost me?

They run all those variables (and more) through a spreadsheet and make their decision based on total cost of ownership.

Bad move.

At this point, they havent even considered the most costly part of the equation—the barrier to exit, or switching cost.

Ive often said that technology has the shelf life of a banana. Call it McNealys Law. It may be a slight exaggeration, but you get the point.

If youve ever heard of Moores Law or Metcalfes Law, you already know just how fast things move in high tech. Whether you buy from Sun Microsystems, IBM or anybody else for that matter, something better will come along in about 18 months.

So, if you stop and think about it, the biggest issue you have is not "How do I get my new IT initiative going?" Its "How, in 18 months, will I be able to kill it? How will I move on to the next-better answer?"

To make my point with a well-known Wall Street company, I said, "Youre a mainframe shop, right? Can you think of any product or technology in any industry—computers, cars, planes, whatever—that has worse price/performance than the mainframe?"

We basically agreed that if the mainframe were an airplane, it would have pedals on it.

"So," I said, "if its that bad, turn off your mainframe tonight and replace it with a faster, smaller, cheaper, more energy-efficient server. Youll lower your electric bill and have computing horsepower to spare." The company couldnt do it. It would have to rewrite all that old COBOL code. Thats a barrier to exit.

The situation is much the same in the desktop arena. The thin-client desktop is way better—its more secure and less costly, has a longer life span, you name it—but people find it hard to turn off their PCs and make the switch. Theyre locked into proprietary file formats, proprietary applications, a proprietary programming environment—all of which create big barriers to exit.

But most buyers dont think about how much those barriers can cost them—or they dont think theres anything they can do about it. Switching is always going to cost them, they figure, no matter which company they choose.

The truth is they can save a lot of money—and change the game—if they ask the right questions.

Is the technology based on industry standards?

Are all protocols and programming interfaces open, published and royalty-free? Have they been adopted by some or all of the computer industry?

Are all the data formats open and freely available? (If not, your current vendor essentially owns your data.)

Is there a community development process (and not just one company) behind the technology?

Is the source (not merely the distribution) open? (The ability to buy the exact same technology from different distributors wont change the fact that youre tied to a sole-source supplier.)

Is a reference implementation, hardware or software, available in an open-source license? (And, by the way, what are the terms of the license?)

The answers indicate, among other things, how hard it will be to move to the next-better answer—and by "hard" I mean "expensive." Dont be short-sighted. Consider the cost of acquisition, operation and exit.

Scott McNealy is chief executive of Sun Microsystems Inc.

/zimages/5/28571.gifFor reader response to this article, click here.

/zimages/5/28571.gifCheck out eWEEK.coms for the latest news, reviews and analysis on IT management from