Comments on: Bubble 2.0

By: mattmaroon

mattmaroon — Sun, 29 Jun 2008 19:32:05 +0000

Well, at some point one more search query necessitates tossing in another server, hiring another sysadmin, etc, and costs 6 figures a year. At some further point one more query necessitates a whole new data center. If their current data center cannot accommodate any more than exactly 1 billion queries a day, number 1 billion and 1 costs a hell of a lot more than 1 cent.

Shouldn't there be an average marginal cost, since the marginal cost fluctuates greatly, being normally very near 0, then on occasion millions of dollars? If so, it's well above the number you guessed.

The problem is that giving away things for free often precludes a sound business model. That was the whole point of my post.

By: Fredrik

Fredrik — Sun, 29 Jun 2008 19:24:52 +0000

Actually, this is a very common misconception. What you're talking about is the average cost, not the marginal cost.

The marginal cost for Google to serve one query, to go from a billion queries a day to a billion and one queries a day is what determines the supply-and-demand curve and in extension the equilibrium price of a search to zero.

Of course increased search volume force Google to have more servers and staff, however, in the context of one additional query, or one additional user, these costs are essentially fixed and does not affect the marginal cost.

The fact that firms are unprofitable only proves that they have not yet found a sound business model.

By: mattmaroon

mattmaroon — Sun, 29 Jun 2008 18:56:04 +0000

But you're wrong. You're greatly underestimating the marginal cost. Marginal costs include all costs which vary with the level of production. Increased search queries force Google to have more servers, more sysadmins, more people working on scaling (scaling is not a one time cost as a lot of people think, it's an ongoing process) etc. Google itself might have that low of a marginal cost, though I'd be surprised, but most sites' are significantly higher.

If it were on the order of some tiny fraction of a cent, sites with large page view numbers like YouTube or Facebook would already be incredibly profitable. Anyone can make much more than $0.00000001 per click with just a Google Ad. The fact that they are unprofitable proves you are wrong.

By: Fredrik

Fredrik — Sun, 29 Jun 2008 10:32:26 +0000

Actually it is. The marginal cost of a single query for Google is zero (bar maybe the $0.00000001 for the extra electricity to emit some photons in a fiber otherwise not emitted) as there practically are no variable costs.

There is a pretty big difference between marginal cost and average cost.

This has nothing do to with why everybody is loosing money. It just means that people need to adjust their business model accordingly. A marginal cost of zero does not mean you cannot make money, just that the price of the actual product (the one with the zero marginal cost) will also be zero - and that revenue must come from somewhere else.

The search business has taken the advertising road and the open-source business has taken the support- and training road. Businesses with zero-marginal-cost-products just has to deal with it.

By: mattmaroon

mattmaroon — Sun, 29 Jun 2008 09:56:42 +0000

But it's not. The marginal cost has to factor in extra servers, bandwidth, staff. It's pretty far from zero, which is why everyone is losing money.

By: Fredrik

Fredrik — Sun, 29 Jun 2008 09:54:38 +0000

The cost isn't zero, but the marginal cost is – and that is what is relevant.

It's basic economics.

By: mattmaroon

mattmaroon — Fri, 20 Jun 2008 20:12:05 +0000

Subscription revenue is pretty awesome.

By: ikirigin

ikirigin — Fri, 20 Jun 2008 20:10:18 +0000

Tipjoy is going to move beyond voluntary payments, and it'll be interesting to see the difference between voluntary donations to a web service, and a subscription service. The penny gap is huge, but you don't really see web services that cost $0.01/month. We're going to try to fill this hole assuming part of the reason is lack of good micropayments providers. We're going to be flexible on whether payment is mandatory for the web service to turn on.

This should highlight what you're talking about with the change from revenue to user count. What % of users of most web services would continue to use them for $1/mo.? What % will sign up to regularly give $1/mo (or year or decade) out of the goodness of their heart?

This has interesting implications for bootstrapping. How many $1/mo subscribers do you need to not die as a startup? Some might claim with elasticity of demand, it should be easier to get 10 $1/mo subscriptions than 1 $10 subscription.

By: cmer

cmer — Fri, 20 Jun 2008 13:20:04 +0000

I absolutely agree with you. Too many people think web development costs nothing. The cost is FAR from 0. Engineers and servers can get very expensive once you get to a certain level.

Everybody's expecting everything to be free nowadays; I don't think it can last very long either.

By: brlewis

brlewis — Fri, 20 Jun 2008 10:45:01 +0000

Try to explain to someone in a bar in Newburyport, Massachusetts that there's free (advertising-supported) transportation available that's quieter and more enjoyable than riding in a smelly taxi that you pay for, and they'll tell you they already know about it.

http://www.newburyportpedicab.com/

Mainstream perception of "free" is changing.