But enough about Faceboom. The more generally interesting business issue highlighted by their story is that Capital Expenditure ("Capex") does matter for web ventures. In fact, it is a mission critical issue, with good software design at the heart of the issue.

Servers Are An Operating Expense For Web 2.0 Ventures

With user generated content, you don't pay people to create content that you use to generate advertising revenues. So your operating costs are R&D (developers), advertising sales and all those senior management overhead lumped into the General & Administrative (G&A) cost bucket. I don't really understand what 1,000 people do at Facebook, but that is another story.

As you scale, people costs should not scale. Servers do need to scale. That is where Facebook must be suffering from some sloppy early software design. That is fine, the initial win is all about user traction and a scalable design is secondary. But today it is a critical issue for Facebook. It is also a critical issue for any venture starting out today. Spending a few bucks early on to get a scalable architecture seems sensible. This is not rocket science, any competent software architect knows how to do this.

Should Servers Be Outsourced Or Leased?

Capex sounds old-fashioned. Why buy servers when you can lease or rent? If Facebook leased rather than bought servers, they could have positive free cash flow even at the lower end of their revenue forecast. The credit crisis will make leasing a bit tougher. Renting via Hardware As A Service (HaaS) is the ideal route for startups you benefit from the scale of the HaaS vendor. But it is unclear how the economics scale for the buyer? It is unclear whether Amazon AWS or any other HaaS is a serious option at Facebook's scale (or the scale of any VC funded venture that is nudging profitability).

Scale Or Profits - The New Choice?

It seems that ventures that can see a fairly quick way to profitability, simply ignore the VC route. The feeling seems to be mutual. VC look at a lot of the businesses that got to profitability and say "too small".

So, you have to choose either big and unprofitable or small and profitable? That does not make sense. If that is true, is this an issue with Web 2.0 models? Some VCs have seen this as a failure of IPO markets, meaning that public market investors won't trust unprofitable ventures promising they will be profitable in future. This "won't get fooled again" view is natural after the Web 1.0 bubble burst.

Trade sales for unprofitable ventures are unlikely to be the solution in the next few years. Not good trade sales at any rate (fire sales are technically trade sales, but they are not a good result). The buyer will be much less willing to fund losses because their investors will be less willing to fund losses for an uncertain period of time. If the venture is close to profitability it does not need to exit and nobody wants to exit in a down market unless they have to. VC have plenty of cash so this is not a financing issue, if the path to profitability is clear.

Maybe profitability for a lot of Web 2.0 ventures is really close? Maybe it just takes longer for Web 2.0 ventures to get to profitable - but that they will be fantastic cash cows when they get there?

Image: mlitty