Home    Files PDF Version

software times™  Files...
Posted to the Gilder forum - April 2, 2001

Moats, Barriers and Switching Costs

I think there is a basic misunderstanding of Gorilla Power. Buffett talks about moats to keep the competition at bay. The more usual term is "barriers to entry." When you are talking Gorilla Power you are talking about something else again. Let me explain. Coke has a big moat, about 100 years of advertising and brainwashing. Anyone can make a new soft drink but to change people's habit is very difficult. Brand names are moats that protect franchises. Building a new world wide underwater fiber optic network is very expensive, and when financing is not forthcoming, very difficult. This is also a kind of moat but here the difficulty lies in the other end of the business equation, at the provider's, not at the customer's end. This is what is usually seen as a barrier to entry. With Gorilla Power things are different once more. To illustrate, once you have installed Wintel boxes in most of your offices, you will want to eliminate most everything else to reduce duplication of effort, such as help desks. Once all your data is in Wintel format, it is prohibitively expensive to change it all to say, Macintosh, even if you were to discover that Mac is better. This particular moat is called "High Switching Costs." It is good to keep these three moats separate:

  • Customer's buying habits (moat)
  • Cost of getting into the business (barrier to entry)
  • Cost of switching a customer (high switching costs)

High switching costs are created mostly by what is known in the Microcosm as "architectures," the Wintel architecture, the Mac architecture, the UNIX architecture. If the architecture is proprietary (Mac and Wintel), and if you win at the tornado, then you have earned gorilla power. UNIX is an architecture as well but AT&T never maintained "proprietary" ownership and Sun stole its thunder. Maybe Linux (a flavor of UNIX) will eat into Sun's business.

There are other ways of creating high switching costs. For example, monster.com, the job search web site, lets you enter a lot of data about yourself including resumes and cover letters. If monster.com offers an easy to use service and a very large assortment of job offers, you will have no reason to switch. Creating a huge database of jobs is a barrier to entry to other job search web sites. Let's suppose that someone does manage it. Now they have to convince you to switch. How much work does it take to duplicate all the stuff you already did for monster.com? The answer to this question determines your switching costs! Web sites have discovered this cost and are using it to advantage. For example, Barnes & Noble has stored my purchasing data: billing address, multiple shipping addresses, credit card number, shipping preferences and so on. If they offer the same price as Amazon and slightly higher than unknown on-line book sellers, I am not going to switch. I know they work and I want to save the hassle of typing in all that data over and over again. Then B&N adds other facilities such as a wish list to reinforce the switching costs. This list let's me store at their web site a list of books I'm interested in for later purchase. It's a lot of work to move this list to say, Amazon. This is the kind of lock in that web sites are trying to create and they do it by letting you personalize the web site. Since you add work, which is the same as cost, you tend to stick there.

Others have mentioned that they fear barriers to entry are limited for PRSF for example - do you agree?

Barriers to entry in software, as I defined them above, are low -- just sit down are write some code! But, can you switch happy customers? Not likely because "architecture" locks them in. How about a new customer who is not locked in to anything? The herd instinct will favor the market leader (you can't lose your job by buying IBM). The important point in evaluating Portal Software is to be aware of the tornado they are in. They are signing up customers right and left including the leaders in many markets (i.e. AOL). Between the herd instinct of the customer base and the switching cost created by the architecture you have a winner!

So if you had one dollar to invest today, would that dollar go toward AVNX or one of the software players?

If it were that simple! Avanex does not have gorilla power but they have Cao and they can put more lambdas on a fiber than anyone else. Cao's switchless network has a problem, the customer has to assemble a "whole product" made up of many different components and the stuff from Avanex is just one of them. Granted, Avanex is the driving force because it is the one creating the large number of multiple lambdas, but you still have to have it all before the network works. The way I see it, Portal and Muse are less dependent on others to crate a whole product and they have been in business longer building their value chains. I would not exclude either Avanex or the good software companies from my portfolio.

But also consider price. The bubble drove prices way too high. The bear market has already compressed Portal's and Avanex's P/E ratios to "reasonable" (whatever that means ;-) levels but Muse is still way too high. Checkpoint Software is rapidly approaching "reasonable" levels.

"Demand creates queues. Supply gets rid of them."

Home    Files Top
Copyright © Software Times, 2000, 2001, 2003. All rights reserved
Last updated June 22, 2003