Cognitive Labs

I read this paper by the Santa Fe Institute's W. Brian Arthur back in 1996. I was just in the early stages of my doctorate and also was product manager of UPS Online. Our team was tasked with figuring out what to do with the world wide web (WWW) as they used to call it. We made up stuff as went along. What was clear was that there was a seachange in processes brought about by collaboration.

In his view, technology was a casino with different tables, multimedia at one, the WWW at another, semiconductors at another. Success was defined by skill in gaming - defining the game, changing the rules, and making sure you understood the strategies of the other players. The article is somewhat relevant today, even though the names might have changed...

Arthur proposes in this paper his theory of increasing returns and he uses the example of DOS vs. CPM, and also, the early growth of Netscape.

The corollary is "what gets ahead, stays ahead" and is seen in the steep adoption curve of such technologies.

Somewhat after this, Google said "if users come, so will the money" referring to experimentation and coming up with something great before the avenues for marketability are readily apparent. The bigger your audience is, the quicker and more relevant the feedback.

Then you go back and alter your product mix based on the user reaction. For example, some tests here get a 10% sign-up rate. Others get 50% Others get 80%!

There are certain characteristics that define the 80% home runs. It's probably not what you'd think.

The user-defined home runs tend to be the simplest offers where the service provider (cognitivelabs.com or other) just gets out of the way.

Sometimes what we think is great in fact, is thoroughly rejected. Other times what we suspect needs improvement in some way, in fact, is exactly what people want (e.g. MySpace)

Below, please see the sigmoid function of the Logistic curve known as an S-curve





The equation defines basically an exponential function. The S-curve is often used to describe the growth of runaway hits. It follows this rule: 1,2,4,8,16,32,64.

This approximates the growth curve of Netscape years ago.

It also defines the growth of an embryo in biology, which divides according to the same formula.


Now I am going to show you the slope of Cognitive Labs growth:




So you see, there is a fairly good correlation between the classic S-curve and the growth of cognitivelabs.com.

// posted by neurofuture6:41 PM permanent link