About the Dialog Decision Process at GM: "The phases of the process are framing, alternative generation, analysis, and synthesis. GM is adept at using this process to guide market research. They know what information about the customer is important for decision making and they get it, quickly and efficiently.

At GM the process was called the “dialog decision process.” Collaborative design  is a better description. Extreme collaboration is required to design a product as complicated as an automobile."

Great DI raw material!  Please add a comment below if you use this resource.  "Academics, or anyone else who’s interested can join the site and start sharing. NASA’s 42 gigabyte map of Mars may be a good start, and a recent copy of Wikipedia might come in handy too"

A great course, from some of the best.  I'm thinking about going.

Watching HANK on Netflix tonight.  Worth a watch.  Another data point validating our intuition of increasing complexity in recent times:

"Along with innovation came complexity.

And complexity is the enemy of transparency.

And complexity is not a good thing in finance."

[24:33]

...

"The domino theory isn't an accurate one here.

The best way to describe it is the popcorn theory.

We had this giant popcorn popper, and it had been heating these kernels for a year as the crisis went on. 

And Lehman may have been the first to pop, and even with Lehman we knew that the same weekend that Lehman was going down we had Merrill Lynch and AIG. So we had three institutions going down. And Merrill Lynch and AIG  weren't going down because of Lehman. They were going down because they had their own bad loans...

That week was like no other week I've ever had.

…

We were juggling difficult decisions, each of which would have demanded 100% of my time at any other time."

[56:00]

Ah, Microsoft friends, you are getting better, but there's one last step that's missing. 

1) "it's not just about offering data sets; it's also about analytic models and other things."

Good so far.

2) "There are fairly well-known models for things like fraud detection, spam detection and such. People are going to be building these models, and we expect to package them with a deployment environment on Azure. For every one person who can build a model, there might be 500 who could deploy it and run it in our cloud environment. Where do they get the data they'll need? That's going to be available on Azure as well. "

Yes, very nice. This is the Market of Models...

3) "We demonstrated one example at a supercomputing conference where we took historical meteorological data over decades and historical airline flight-delay data over decades and we built a predictive model that combined them. We could then ask, "On a clear day, what are the probabilities of delay for various airlines, airports and times?" Based on the historical meteorological data, we could also ask, "What do those probabilities look like when there's six inches of snow in Detroit?" If you then add meteorological forecast data, you can ask, "What are the chances I'll miss a 30-minute-window connection in Detroit next Tuesday?" "

Getting closer, you're starting to connect the dots. 

*But*...

The question is not "what are the chances I'll miss my connection", but "what flight should I book to a) minimize my chances of making it; b) make it most likely I can fly on a 787; c) minimize my miles in the air; d) avoid Detroit if at all possible?"  Microsoft, you need to ask the *decision* questions, and connect *all* the dots to include the outcomes as well as the full decision model chain. 

Call me, friends, let's talk.

A landmark paper and the most accessible treatment I've seen on why more data does not produce better decisions.

"Consider a block sliding down an inclined plane. To address the micro scale—the molecules—we average over them and, using thermodynamics, describe their temperature and pressure. To address the macro scale—the block and the inclined plane—we use Newtonian physics to talk about their large scale motion (see Fig. 2). In this case, the pieces can be considered to be acting either independently (the random relative motion on the micro scale) or coherently (the average motion on the macro scale) and since the scales ares ufficiently distinct we do not encounter a problem in describing them separately.

But many systems, especially those we are interested in understanding and influencing, are not well described by two separate levels. Consider a flock of birds as it leaves a tree in response to a crack of thunder. If all of the birds flew independently in different directions, we would need to describe each one separately. If they instead all went in the same direction, we could simply describe their average motion. On the other hand, if we were interested in how the birds move from the initial mostly independent directions of movement, into groups and then came together as a flock, describing each bird’s motion would be too much information and describing the average would be too little information."

This is the press release. The paper is here: http://www.necsi.edu/projects/yaneer/information/infoimportance.pdf

Why is this relevant to decision intelligence?  Because, according to Steve (imho one of the smartest DI guys around), effective decision making is a key competency as your startup grows through the 10 people / $5M in revenue mark.