Just a few years ago, in the middle of the real estate bubble, there was mounting pressure on McDonalds to break apart their business.  The rationale was that McDonalds real estate – the physical land that the restaurants sit on – was worth more than the profitable operating company that employs 400,000 people and makes hamburgers.

I won’t comment on the social ramifications of the plan, but many smart people thought it was a good idea.  There was only one problem behind this logic.  McDonalds property is zoned for restaurant use, and no other restaurant in the world extracts more profit per square foot than McDonalds.  A rational observer would conclude that in order for this idea to work, the value of the land needed to exceed the maximum value that could be extracted from it.

In retrospect this was a clear sign that we were in the middle of a real estate bubble.  Underlying asset prices were untethered from reality.

The last Internet bubble in the 1990s had similar dynamics, although for Internet real estate.  Investors speculated on the value of “eye balls” (code name for visitors to websites) and they poured money into consumer properties based on the number of users, without any regard to how those users translated into revenue.

In the last week Facebook published figures showing that they are now generating about $4/user/year.  I found a chart below that shows comparable math of how other major web properties monetize traffic:

(Source: Silicon Alley Insider)

Before you jump to the conclusion that Amazon is the best company ever, keep in mind that not all revenue is created equal and there is a difference between retail and media businesses.  If you just focus on the ad-supported businesses to the right of eBay, Google does a better job than anyone else of turning Internet real estate into cash.

What does this mean?  A couple thoughts:

1.) Google can afford to pay significantly more for web properties than its competitors because it makes more money from the traffic.

2.) Advertisers pay more for reach.  This means that big platforms get a multiplier effect by combining higher traffic with more revenue per user.

3.) A reasonable revenue target for most ad-supported start-ups is $1/unique user/year.  It is hard to scale beyond that unless you have highly engaged users (ie lots of page views for each unique) or you can invent your own advertising model.

Early stage investors have always been focused on traffic because if you don’t have users than nothing else matters.  But I am surprised at how many late stage investors are taking monetization risk.  The venture industry should be careful about that approach because it didn’t work out so well last time around – at some point somebody has to make hamburgers.