What early adopters really do at their computers 

TechCrunch has a great post on how early adopters spend their time on their computers. What’s fascinating about this is the time usage stats are based on actual behavioral data, so it’s very accurate, relatively speaking.

No doubt people are drawing all kinds of conclusions from this. I thought I’d share mine, which revolve around how this data affects market share, revenue, and valuations:

  • Gmail is used 3x more than Google.com. Not surprising when one considers their own workday activities. But, assuming clickthroughs are more or less equal for both (valid assumption?) — that means Gmail generates the bulk of Google’s Adwords revenue. Pretty amazing considering that Gmail originally came out of a developers "20% time" project. This supports the notion that sometimes the best projects come out of skunkworks.
  • Facebook is accessed 50% more than Google.com. Maybe that $15 Bn valuation is justified after all!
  • Outlook is, unsurprisingly, the most used app. Now, think about Xobni. If you got it installed, whenever you use Outlook, you’re using Xobni. That means Xobni could soon become one of the most widely used apps around. That presents some interesting monetization opportunities when you have that many user attention minutes. It will be interesting to see what the future holds for those guys.
  • It’s surprising that Yahoo Messenger has such low usage. Last time I’d looked, a few years ago, Yahoo had many instant messenging users. One more thing for Jerry Yang to worry about. I’m also surprised that Meebo is at the bottom of the list.
  • OpenOffice and Google Apps have very low usage. For instance, Google Docs has 3% the usage of Word. If even the early adopters aren’t using them, I guess it will be some time before they start to challenge Microsoft in terms of market share. And it will be some time before Microsoft profitability, largely driven by the Office suite, starts to suffer and drag down Microsoft’s valuation.

A note on accuracy: I’m sure some will quibble about the accuracy of the numbers given that
the sample was self-selected, But market research is an inexact science.
Not to go all Rumsfeld, but you need to use the data you have, not the
data you wish you had. This, as far as I know, is the best data we have on what people actually do on their computers. (If you know of a better data source, please let me know.)