Profit per Byte

I have this thing for Jim Collins- the guy who co-authored Built to Last and Good to Great. It's a little unusual, and I don't like to talk about it- my fellow engineers sort of look at me funny when I talk about business stuff, but I can't help it. I find it all really fascinating. We have all of these buzzwords and frameworks and processes for people who want to make business seem more interesting than it really is, but I've never felt like that stuff was required to make business interesting. That's not to say that Collins' isn't guilty of this sometimes, but he's better than most.

One of the things he talks about in Good to Great is the "Hedgehog Concept", which is modeled after some fable about a fox and a hedgehog, or some such thing. The real idea underneath the packaging is a focused understanding of what drives a company's bottom line, and how that understanding drives everything a company does. Collins argues in G2G that one way to really get at this core concept is to think about a core profitability metric- something like profit per x. For some businesses, this is profit per widget, and for other (more service-oriented) businesses, it is profit per employee. Walgreen's is a neat example, it's metric is profit per customer visit. The book goes into all of this stuff in more detail, but I wanted to focus on the web, and compare two different profitability metrics that I think are in play: profit per pageview, and profit per byte.

The profit per pageview model is what drives the vast majority of websites today. The basic logic seems to be that whenever a user should chance upon a page within your site, you should throw as many ads at them as you can possibly squeeze onto the page, in the hopes that something might get a click from the user. It is, with differing degrees of subtlety, the model that drives Yahoo!, eBay, Amazon, and most of the other old-school Internet properties.

Google, on the other hand, has to be the prime example of the profit per byte (PPB) model. Maximizing PPB has a number of implications, which can be seen throughout the Google empire:

• Lightweight webpages that avoid big graphics and complex HTML,
• Text ads that take up only a few bytes, as opposed to animated banner ads,
  
• AJAX-based websites (like Gmail, Google Maps) that don't require the user to download the same bytes over and over again- only the information that changes

It's interesting to me that while so many companies have started copying Google's look and feel, their computing infrastructure, their work practices, etc., there hasn't been much discussion of copying the business model that drives so many of these decisions.