Google's real problem: the smartest kids in the room

For a phenomenally successful company, Google makes a lot of mistakes. Google TV, Buzz, Wave, privacy infringements that earned the company 20 years of privacy oversight - and the whole one-trick pony thing, with 97% of Google's money coming from ads in search.

It's too early to call Chromebooks a failure, although I'm tempted. Then there was losing the Novell and Nortel patent portfolios, complete with nose-thumbingly clever bids, complaining about it - and forgetting to mention that it turned down a partnership with Microsoft because it didn't fit with its plans to use the patents (offensively or defensively). Continuing arguments about whether blocking Google+ users with pseudonyms rather than what Google thinks sound like real names (and in some cases blocking them from other Google services including Gmail) are obscuring the possibility that Google+ is going the way of Buzz and Orkut because you can't write an algorithm that understand social behaviour.

And the success of the Motorola acquisition is far from certain, what with the show-trial-identical statements of support from the partners Googorola will compete with, the increasing likelihood that Microsoft will win its case against Motorola at the ITC and the charming welcome some Googlers are extending to their new hardware colleagues. A report in the Wall Street Journal briefly included a quote (that now seems to have been removed) from a "former Android executive" that Google employees have 20 more IQ points than Motorola folk.

I winced when I read that because it sums up perfectly what I think is Google's biggest problem. All their lives, the Googlers have always been the smartest kids in the room - and whatever room they walk into, they assume they still are. That might often be true; many of the industries Google has disrupted have been full of deadwood and dinosaurs living off past triumphs. But assuming that you're smarter than your competition, and that what they do is easy enough for you to replace it with an algorithm, can cause problems when you underestimate them.

When Google first came out with its own phone that you bought directly from Google, got support on in a Web forum and put on a service plan from a Google database, there were hints that Google wanted to disintermediate those pesky mobile operators and solve the impenetrable tangle of mobile phone plans by creating an algorithm to pick the right one for you. How hard could it be?

Given that mobile operators haven't usually finished introducing the last set of phone contracts when they start advertising the new ones to compete with what another carrier just announced, quite hard. At the next Google IO, it turned out that just keeping the database of different phone plans up to date had been too much work and Google starting calling it a research project to help them understand carriers better, because they'd be the ones selling Android phones.

Google not offering handset makers an indemnity against getting sued for any software patents their Android phones infringe when there are thousands of IP licences covering the hardware side of phones? Being isolated from how much phone IP there is because they weren't making the phones themselves might have helped Google underestimate the strength of software patent portfolios.

Bidding Pi billion for the Nortel patents? It's one way of making sure you don't waste a bidding round by coming up with exactly the same offer as an opponent, but it also smacks of 'look how smart and geeky we are'.

As Intel's Andy Grove used to put it, "only the paranoid survive". Assuming that all the smart people work for you is really dumb. Mary