Skip to content →

Tag: forecasts

Don’t count your markets before they hatch

I was reading a very insightful analysis of the supercomputing industry over the past decade on, when I stumbled on a chart which illustrates not only a pattern I see very often, but also a reason why you should always sandbag your forecasts if you’re betting on a new technology: your forecasts are almost always too optimistic.

Take Intel’s and HP’s huge gambit to push Itanium as the processor technology which would eventually replace all the other major processor technologies (i.e. SPARC, PowerPC, even Intel’s very own x86). Countless technology analysts and Intel/HP market researchers said Itanium would become the only game in town in computer processor technology – and with good reason. The technology that Itanium represented, in theory would’ve completely changed the processor technology game.

Yet, if we look at the progression of Itanium sales versus Itanium forecasts, we see a very different picture:


If anything, Intel/HP have now distanced themselves from Itanium – preferring to ship products based on Intel’s homegrown x86 technology, rather than the technology analysts had expected to storm the market.

Kind of embarrassing, isn’t it? The point isn’t to call out HP and Intel’s folly here (however much they deserve it), but to point out that new technologies are notoriously hard to predict, and to whatever extent is possible, companies everywhere should (a) never bet the farm on them and (b) watch what forecasts they’re making. They may come back to haunt them.

(Image credit)

Leave a Comment

My Hobby

It’s not just Dilbert who makes fun of forecasting, XKCD is into it as well:


I wouldn’t call it a hobby, though. More like a day job that involves doing what is effectively described above, but making slightly more reasonable assumptions :-).

Leave a Comment