GigaOm discovers Web Growth costs money - Shock Horror!

Henry, the problem is that my bet is that Twitter's expenditure isn't anything like 30 cents per user.

An apt comparison is Facebook. In a call (supposedly private) last year, Zuckerberg put the company's capital expenditure for 2008 at $200 million, with 37 million users. Even if you assume that the company will have no additional capex over the next three years, that effectively means the company needs to make around $2 per user per year to cover it. And that capex is just servers - the bandwidth bill, increased staffing costs, and every other expenditure isn't included.

Now of course that cost per user will come down as you gain users, which, it's safe to assume, Facebook will. But to come down to the "almost nothing" levels you're talking about, Facebook would need to grow to around 350 million users - with no additional server/bandwidth costs. That ain't happening.

In the same call, Zuckerberg also said that revenues would be between $300-350 million for 2008. However, factor everything in and this means that the company would still burn through $150 million of its investors money that year.

Now Twitter's costs won't be as high as Facebook's. It doesn't host images, for one thing. But it's revenues, at the moment, are a big fat zero - and it's difficult to see a business model which allows it to make dollars per user per year, particularly in a downturn. My bet is that it will simply become a search engine, selling ads against Twitter search results.

Yes, they could just charge people - but converting 1% of people into paying customers will be a tough ask for Twitter, without degrading the experience for non-paying customers. If I can still send unlimited tweets to unlimited numbers of followers, why would I upgrade?

You also have to be careful about saying that these kinds of companies are profitable. Lots of them announce EBITDA numbers which on the face of it put them in profit - but don't include a lot of expenses, and, in particular, don't include investment they must make if they're to reach the number of users they claim they need to make a profit.

Facebook being profitable next year depends on its growth in revenues exceeding the growth in new users - and the problem is that this just isn't happening.

Remember that, in server/bandwidth costs, beyond a certain point there isn't much economy of scale to be had - serving 300 million users really will cost you ten times as much as serving 30 million, unless the time that the average user spends on your site decreases. And if that happens, the value of your users to advertisers also drops - which reduces your (potential) revenue per user.

Unfortunately, too, for social networks like Facebook and Twitter, the more users you have the more likely it is that people will come back more often. If only ten of your friends use Facebook, you won't go back to it very often. If it's 100, I'll visit regularly. And if everyone I know is there, I'll be on it constantly.

This means that, unless you have a turkey of a service, more users will also equal higher per-user usage. So, not only do you not get economies of scale as you grow, each user will actually cost you more. Ouch.