Market forecast bunkum (first appeared in InfoWorld)

I have good news, great news, bad news, and good news.

The American Cocoa Research Institute (ACRI) recently reported research on (what else?) the health benefits of cocoa and chocolate.

The good news: The ACRI exists. Something this important deserves a research institute. The great news is that cocoa and chocolate are chock-full of flavonoids. Flavonoids are naturally occurring anti-oxidents, which means they help clean up nasty stuff inside you. According to the ACRI’s research, the flavonoids in cocoa and chocolate are more potent than those in red wine, and in fact are more potent than Vitamin C.

While the current marketplace for flavonoid detectors is only $10 million annually, the Froschbosch Group predicts that by 2005, it will grow to over $2 billion. This prediction is neither more nor less reliable than any of the other triple-digit-growth-rate market predictions you read about every week in this and other IT industry publications.

Have you ever taken a close look at these puppies? They’re all identical. They’re exponential growth curves — not one is S-shaped, none have plateaus, interruptions, or inflection points in their smooth takeoffs to glory. They all start with sales statistically indistinguishable from zero, and end a zillion times bigger. Add together a complete set and the 2005 total is 37 times the current US gross national product.

That’s a lot of economic growth … which, if it meant anything, would be great news.

Sadly, it doesn’t. That’s the bad news. These market forecasts are almost perfectly unreliable.

The statistics themselves are fine. I’m confident that the samples are big enough, stratified enough, stationary enough, and unbiased enough, and the computations are handled with precision. But they’re based on surveys.

So. The Froschbosch Group calls a CIO and asks, “Are you currently using XML to define workflow metadata?” The CIO answers that he is not with confidence.

“Are you actively investigating its potential?” The CIO doesn’t want to look like an incompetent bowb, so “Yes,” he responds, with no real understanding of what workflow metadata is.

“Will it be in production by 2002?” The CIO just said his organization is actively investigating this mystery cure, and two years is a long time. “Yes.” Heck, for all he knows, he will.

But probably, he won’t. He hasn’t even started the internal selling needed to get funding approved. In the unlikely event that he sells it and the business buys it, about three of every four IT projects fails, which means XML-based workflow metadata never will go into production in his company.

Okay, let’s give our survey respondent a break. Maybe the question is about the future of object/relational databases, and the CIO understands the subject. Every RDBM vendor has announced object/relational extensions, and his policy is to stay current on releases. Will he have object/relational database technology in production by 2003? “Yes,” he answers, and the survey company announces exponential growth for the object/relational marketplace.

The bad news is that most of these market forecasts don’t mean very much. What’s the good news?

The good news is that none of them matter to you. Amazing market growth doesn’t mean a technology is important to you, any more than a more modest market growth rate means it’s not.

All that matters is whether your business can benefit from a technology. No growth curve can tell you that. Market growth matters only to venture capitalists and Wall Street analysts, not CIOs and CTOs.