Standards: Bottom-up, Top-down, and other?

I think I’m noticing something I’d not noted before about the statistics of the populations served by the standard. Some standards don’t scale the way you might expect them to. For example the B2B standards written in the 1980s and 1990s have not scaled. They are not used by small businesses. Bear with me as I build some substrate.

Exchange standards provide efficiencies for the transactions between parties. If we all adopt a similar handshake then we can have more handshakes at lower cost. Next up world peace.

In thinking about standards you can focus down on the details of the single exchange, but I find it more fascinating to shift up and think about the populations involved with the standards. For example consider selling a house. Down in the details you can pick apart the process steps taken by a buyer, seller, and middleman they do the transaction. Stepping back, at the real estate market level, look consider the three populations: house buyers, home sellers, and real estate agents.

Statistics gives us tools to talk about these populations. Even the simple statistics can illuminate some interesting things. For example we know that small populations are easier to organize and coordinate. In our example the smallest population is the real estate agents. When society comes to negotiate the rules (aka standards) for home sales the agents are much more likely to get their desires fufilled because they have an easier time getting their act together. The smallest of the three populations in any standard setting scenario have an advantage. That is a political reality.

Another simple statistic: some members of the population do more exchanges than others. The surprising fact is how skewed that is usually is. Some members do a lot more exchanges than others. Again and again when you look at the populations around these exchange standards you find a power-law distribution.

A cartoon approximation of a power law curve splits the population into two groups: the elites and the masses. In the blogging world they call the elite bloggers the A-list. (Sidebar about the risk of cartoons. This crude approximation is blind to the middle class. Blindness can do harm and so can approximations. That said, we return to the fun of this cartoon.)

The small size of the elite population has a political consequences exactly like the small population of realtors. The large transaction elite have power. For example, returning to the rules around real estate sales again, we can look at the population of house sellers for a high volume player – i.e. real estate developers. Not surprisingly they show up at the table to help set the rules.

Ok, so back to the to seemingly new thing.

What I’m noticing today is that there seems to be some very interesting to say about what happens about the correlation between two population statistics.

What happens if the elites tend mostly to exchange with only with each other? That’s what happened with the B2B standards that were written in the 1980s and 1990s. Large businesses recognized the benefits of getting standards in place to improve their efficiency. So they wrote all these standards to fit their needs. Today lots and lots of commerce takes place intermediated by those standards; but the majority of that commerce doesn’t involve small economic entities.

It’s probably worth ringing all the changes here, but that’s a project for another day. For example you get a particular kind of standard when there is a small concentrated elite on one side exchanging with a huge diffuse lower-class on the other.

If a standard is designed with only one class of players, large transactors say, then it won’t be well suited to the needs of the players who were not there when it was designed. I think that’s a very interesting insight.

Let’s go back to the example of the B2B standards designed in the 1980s and 1990s. Consider this design question: “How hard/expensive should it be to adopt this standard we are designing.” The answer the elites gave was “No more than two expert consultants for 3 months.” The answer the masses would have given: “Oh, $49.95 would be acceptible.” There is a third group at the table. The guys that design and implement the standards, i.e. the vendors. Their answer is always “As much as possible.” So the B2B standards of that era got designed with a high adoption cost, i.e. they are standards with very high barrier to entry.

This doesn’t always happen. Sometimes a standard comes out of the masses. As SMTP or HTTP did. Small players solving a problem that then got widely adopted. Such standards have their own problems; for example they may not scale well for the traffic patterns that the elite players experience. If you fear the power-law’s tendency to concentrate power you might like this kind of standard. If your trying to consolidate a market you might prefer the other kind of standard making.

Leave a Reply

Your email address will not be published. Required fields are marked *