I have had a quick look at the Facebook prospectus.  The thing I found really interesting was not confirmation of revenue and its dependence on advertising, but a very first glimpse of how Facebook’s costs are structured.  I  believe the key to working out a long-term, sustainable valuation for Facebook as a business lies in its costs – for the very simple and old-fashioned reason that a business cannot create marginal revenues that are significantly greater than marginal costs (if it is operating in a functioning, competitive market – which of course Facebook is not at the moment).  If you therefore want to understand long-term revenue potential, understand marginal costs.

Facebook has shown its costs thus:

• Costs of revenue: $860 million
• Marketing and sales: $427 million
• Research and development: $388 million
• General administrative:$280 million

This, of itself, is quite an intersting breakdown.  What other companies might there be where marketing, sales and R&D account for 41% of total cost?  A pharmaceuticals company perhaps?  However, what Facebook hasn’t done is split costs out in a way which allows us to see how much Facebook is spending on people, versus how much is spend on tangible operation costs, such as server space and data centres.  What Facebook labels “salaries, benefits and share-based compensation” is included within the figures for each of these categories, including the ‘Costs of Revenue’ line.  One would suspect that, in reality, people costs are a very sigificant element in all of these categories and it would be nice to know how this breaks down – not least because it will give us a glimpse of what it actually costs to deliver and maintain the infrastructure that is Facebook.

Knowing this is pretty important because it allows us to get a handle of what Facebook’s true marginal costs are and thus its true revenue potential.  What is already clear is that most of the costs associated with marketing, sales and R&D are not true costs, in that they are the costs Facebook has elected to incur in order to support the business model it has chosen to construct.  They are discretionary costs, in that any competitor coming into the market need not incur them in order to deliver a similar service to users.  In effect, Facebook has decided to incur these costs in order to service advertisers and chase sufficient revenue to justify a valuation in excess of $60 billion, rather than the revenue it needs to actually deliver the service to its users.

Look at it another way.  Facebook has 845 million active users.  Match this against costs of $1,955 million and this comes out at $2.30 per user.  Now I know this isn’t a truly accurate way of establishing marginal costs – but it probably isn’t way short of the mark.  So are we to believe that it costs Facebook around $2 to service every additional user.  Of course not, it will cost Facebook much less than this.

If I had to take a stab at it, I would reckon you could deliver A Facebook (if not The Facebook) to users for something in the region of $300 million.  Take that, plus a margin of say 25% and that gives you revenue of $375 million.  Put a multiple of say 15x on that and you get a valuation of $5.6 billion.  The fact that Facebook is already earning revenues of $1.7 billion, and carries a valuation in excess of $60 billion might give one pause for thought.  I don’t earn millions as an analysts at Goldman Sachs – so I wouldn’t necessarily set much store on my opinion.  But if I were an analyst at Goldman Sachs, I would certainly want a whole lot more information on Facebook’s costs.