Monday, September 27, 2004

Microsoft, Competition, Market Structure

by Asim Jalis

I had a really good weekend. It all started when I read through an interview of Bill Gates by David Allison of the Smithsonian Institute. I had read this a while back, and then stumbled on this again on Saturday. The interview hints at principles, but does not lay them out. Let's start from the beginning. Suppose you want to start a venture. Let's say you want to become a successful shareware author, or you want to become a bestselling writer, or a highly successful blogger. The question arises: What should you write about or sell? The difference in degree of success can be dramatic. Compare the profits of a blog gets 100,000 page views per day versus one that gets 1,000. This number is profoundly significant. It determines the scalability of your effort. Here is the answer (or an answer): Instead of reasoning through the mind of the market to figure out what will sell, look at what is selling, try to understand why, and then create a competing product. We've discussed this idea before but not in this way. The competing product does not have to be identical. It can be different in small ways, but it must address the same basic problem. Given that we have limited information about the market, all information is useful. The fact that someone is already succeeding is a piece of information. Using this gives you more leverage than if you ignore this feedback. This is a lower-risk strategy than creating a completely new market. The innovative strategy can work really well, but it requires patience and many experiments. This strategy is likely to succeed with less investment and might be a good one for someone with limited capital who is trying to start his first venture. Let me make this more concrete. Let's look at, which might be an example of a highly successful blog. Why does it succeed? Here are some possible reasons: 1. It's possible that while the number of PVR users in the world is very small, the overlap between the people who use PVRs and who surf the web for information is extremely high. 2. The web is good for reviews of innovative stuff. (a) There is not enough information about new things. People are always looking. (b) The people who like new things are also likely to use the web a lot. 3. PVRs lead to high-paying Google Ads. So now, what if someone else enters the market and creates a me-too PVR blog. Even if this blog gets 10% of the revenue of the main one that might be several fold more than what it could get if it was about puppies or grits. It might even be possible to get more than 10% of the market share. For example, the competing blog could improve on the original. It could have a Q&A column where people can ask questions. It could have contests where it gives away PVR accessories. There are tons of things that could be experimented with to make it better. The reason these experiments are feasible is because we have narrowed the search space. Competing narrows the search space which allows useful experiments to be run. Competing creates a structure in the market. Otherwise reality is unstructured and it is hard to figure out what to do and where to move. Notice the dialog between the Republicans and Democrats. They frame the conversation for each other. In a one party state there would be no structure to the conversation. E.g. in most companies where upper management runs the organization like a one-party state it's hard to get engaged with issues because things are unstructured and unframed. Life become interesting when you have competitors. Microsoft stays engaged precisely by having competitors. Competitors turn work in play. Suddenly it is a game, and you ahve to win. It's much more fun to play with someone else than to play alone. Competition narrows and structures the search space. Let's now turn to shareware. Looking at, the bestselling software is all for cleaning up pop-ups and spyware. This category seems to do really well. While it makes sense that it would do well, it is not obvious why it would do so much better than other useful categories of software. And yet for whatever reason, it is big. So if I write a trading game and completely corner the market, I am going to make much less money than if I write a spyware cleaner and capture only 1% of the market. Now it's possible that a trading game market could eventually also become the same size. However, for a category to succeed many things have to line up. Many of these things are impossible to observe or even understand. For example, it's possible that anti-spyware sells so well because companies buy it to install on their internal networks. Companies are more likely to invest time and money in anti-spyware software than in other useful software, because the danger from spyware or viruses to an IT network are so serious. All of this reasoning is completely speculative. The main point is that for a category to be successful a whole ecology must exist around it. And it is nearly impossible to create such an ecology from scratch, or to foresee it before it springs into existence. One of the reason both Britain and France became world powers is because they were right next to each other and spurred each other's improvement. This is the same reason that all the dotcoms are in California. It's much easier to compete with people across the street than half way across the world. And competition makes everyone stronger.