Microsoft and the Magic Coffee Machine

Posted on June 23, 2003

Lots of buzz in the blogosphere this morning over Eric Kidd’s essay, “The Missing Future“, which asks whether there’s any future for small, independent for-profit software companies as they get squeezed between Microsoft on one end and open-source software on the other.

Perhaps the strongest opposition to the piece has come from Microsoft evangelist Robert Scoble, who argues (surprise!) that it is possible to enter a market dominated by an entrenched monopolist and carve out a profitable niche nonetheless. He illustrates his thesis with the examples of a small Seattle coffee shop called Victor’s that does well even in the home of Starbucks, and an ice cream chain that thrives in a world dominated by Baskin-Robbins.

Scoble’s piece is interesting, but I think he misses the point. Imagine if there were a technology that could make cups of coffee appear magically on your desk, whenever you wanted one. Imagine that, say, 95% of the world got all the coffee they ever drink through this magic coffee machine in the sky.

Now imagine that said technology by default would only serve coffee from Starbucks. You could get coffee from other vendors, but you’d need to fill out a bunch of forms and wait six to eight weeks for processing, and even then the other coffee would never be delivered perfectly; every fifth cup or so would materialize only half-full, or tasting like drain cleaner, or spilled in your lap. The Starbucks coffee, though, would materialize perfectly every time.

Would Victor’s Coffee be able to compete in a world like that? How many people would jump through the flaming hoops to get a cup of their coffee through the magic coffee machine instead of just accepting whatever Starbucks gives them? Even if Victor’s coffee was head and shoulders better than what Starbucks handed out, most people would figure that the Starbucks coffee was good enough to justify not going to the extra effort to get Victor’s.

This is the world that independent software developers face today. The magic coffee machine is Windows, and Microsoft is Starbucks. You can go to the extra effort to try to brew something better than what they’re brewing, but at the end of the day you still have to serve your coffee through the magic coffee machine to reach the 95% of people who are locked up inside it — and since Microsoft owns the magic coffee machine, they have advantages in delivering coffee through it that you don’t have. You can be positive and say that their advantage is that they know its innards better than you do, or negative and say that they can add speed bumps to keep anybody else from using it as well as they do (“DOS isn’t done until Lotus won’t run“), but either way the effect is the same: every fifth cup of your coffee ends up in somebody’s lap. Open-source products can survive that because people expect them to be unpolished and clunky — hey, it’s free, what do you expect? — but if you’re charging good money for your product people are going to start getting upset the fourth or fifth time their morning joe tastes like drain cleaner. Only the most motivated are going to stick with you through that, and most people just aren’t that motivated.

That’s what Scoble is missing when he critiques Kidd by telling him that if he had a good enough idea he could compete with Microsoft: no matter how good Kidd’s coffee is, it doesn’t matter when he’s competing against the guy who owns the magic coffee machine.