In growing markets new buyers lack information to select the best vendor, the one that fits their needs best. In this absence of information they grasp at straws; other measures which they can understand. Proxies for quality. The simplest model for why you get highly skew’d distributions, like those seen in market share numbers, is to have new entrants link to whom ever already has a lot of links. Market share in one time period generates market share in the next; not just a little bit, but a lot! This is why early movers can have so strong an advantage. They can then create a brand – and a brand is nothing if not a proxy for doing the real work checking if a vendor fits your needs.

GapingVoid is yet another blog written by a consultant who’s attempting to puzzle out how the Internet is reshaping his craft. He’s a marketing consultant. For example he’s currently experimenting with liqouring up gatherings of bloggers to see if that rebounds to the benefit of one of his clients, a vineyard. One of his success stories is English Cut, a high end tailor – $4K/suit. This posting muses about how successful that experiment has been. It’s kind of guilty gloating. He quotes an observer who notes how the English Cut has captured the hub for high end tailoring. It’s become the brand in high end tailoring. He also muses that this kind of tailoring isn’t a particularly scalable business; double your customers doubles your work. Which is trouble if you have to do it all by hand yourself.

All this is very mysterious. Clearly blogs inject more information into the market. Since lack of data aids early movers blogs would appear to tempering early the mover advantage of the mindless linking to what ever is popular. But just as clearly blogs don’t do that. One reason they don’t: the blog gets caught up in it’s own early mover effect. Observe any class of blogs and you’ll find few hubs that arrived late to the game. Worse yet the late arriving hubs often appear to be explained by their association with older or larger hub. That skewed distribution means that if your a winner then it’s Ker. Ching. and if your not then your a dead body.

This seems to it implies that the globalization implicit in blogging is very damaging for small players. If the market doesn’t expand tremendously then blogging is just another tool for consolidating an industry. That blogging, in this case, is just another distribution channel – a distribution channel of marcom. Distribution channels define the links, The links lead to the skewed distribution. The skewed distribution creates a pile of dead bodies. Traditionally some industries, like high-end tailoring, because they don’t scale the participants don’t threaten each other and the members can be quite collegial. Enter the internet where the network effects rule and that collegiality at risk. It’s winner take all.

Paul Ewald’s book is a rant against conventional wisdom. It opens with a flat out denial: parasites and diseases do not tend to evolve toward more benign relationships with their hosts. The conventional wisdom is based a series of just so stories, an optomism that would do Pangloss proud and a Kansas school board model of evolution.

This stuff has consequences. It’s actionable. Bacterial, viruses, etc. evolve very quickly. When we change their environment they are almost always find ways to leverage those changes. Decisions about public health can leverage that in positive ways. Or, they can blindly ignore it and creating horrible unintended consquences.

I read, rather than skimmed, the whole book because the stories are rich in analogies to the stories I’m most interested in: those about middlemen, platforms, and networks. For example mosquitos act as a intermediary for malaria. In this domain they are called the “vector.” Like the postal system, UPS, or the rail roads they provide transportation services. Understanding many of the stories in this book demands picking apart how how the infecious agent evolves in the face of pressure from both the distribution channel and the host. If the nature of one changes the bacteria (virus, tape worm, etc.) changes to strike a different balance.

There is a facinating insight here: severity of the disease is tied to the nature of the distribution channel. For example diseases which use mosquitos for as their vector, like malaria, are usually more severe than airborne diseases. A mosquito borne disease tends to be severe so it can immobilize it’s host and assure that the mosquito has an easy time feeding and when it does it’s meal contains carries the infection. If we can arrange for the patients to be moved into well screened houses where they can’t be reached by the mosquitos then this scheme falls apart. In which case it’s preferable for the affliction to evolve to keep their hosts mobile – i.e. a less severe varient of the disease emerges. This kind of modeling suggests why the common cold is relatively benign (it needs to keep the host mobile). It is very suggestive about why the trenches and field hospitals of the first world war may have generated the 1918 influensia epidemic; where the army provided continual supply of fresh hosts and mixed them intimately with those infected.

Key to many of the interesting scenarios around networks, standard, and businesses are the situations where the links are made between two different groups and these stories with the disease is the leveraging services of two distince parties seem quite analagous. A middleman, in a business context, covers his expenses by charging the parties on either side, usually differing amounts. So the dating service will charge men more than women while eBay will charge sellers while it advertises to buyers.

The same pattern happens here. Malaria is reasonably benign from the point of view of the mosquito. Analagously the fraud around eBay, the broken hearts around dating services, the viruses on Windows, the spam in your mail box are all reasonably benign from the point of view the intermediaries.

Reading this book you begin to think that any time you see mixing between two classes of actors you need only look and you’ll find a parasite that’s discovered a way to play the middleman. The stories I found the most disturbing are the ones where caretakers become the vector. There is a disease of coconut palms that uses the machette’s of the plantation workers as it’s vector. That story has the horrible plot twist that there are two ethnic groups and only one of these group’s plantations were infected. It had nothing to do with how they ran the plantations, only that the disease agent was issolated because the two groups never exchanged machettes.

He believes, but doesn’t quite have the research to prove, that many of the horribly virilent diseases that have emerged in hospitals over the last few decades can be explained, and then controled, by using these ideas. That these deseases have evolved so they can use the doctors and nurses as vectors and the patients as hosts. The key to pulling that off is to evolve to be benign in the vector and virilent in the host. Any difference between patient and caretaker is an opportunity waiting for a mutation to leverage. Newborns are particularly rich in these differences. So are patients taking antibiotics because they have suppressed their entire spectrum of bacteria. There is an ugly story about an outbreak of murderous diarrhea in Chicago. All cases were traced back to 27 hospitals; but how did they spread between the hospitals?

So, are the end to end principle (pdf) and worse is better really the exact same idea? Both are kind of negative in tone. Both have MIT appearing as a character in their stories.

End to end isn’t a paper about the risks of agency, or middlement, but those issues are clearly in the background. Worse is better is more conscous of the social engineering that is going on as the designer makes choices about the shape of his system. Neither is particularly clear that the designer is crafting an option space, a search paridigm, a platform, or a standard. To my mind both have become deeply assocated, over time, with our culture’s romatic notions about the little guy, the rural, the entrepenur. Ideas that are currently appear in the role of “long tail.”