1) Randomized A/B testing is the way to go for understanding what works and doesn’t. Examples: Offermatica, CapitalOne offers (for optimizing what wording/colors will get you to accept an offer), Brooks Bell Interactive (for optimizing your website conversions)
2) Owning data, and putting it together in a powerful connected way seems to be both a short and long-term competitive advantage. Facebook, Apple, Amazon, and Google will be hard to unseat, partly for this reason.
Acxiom, an example of a company that owns a lot of data and sells it
3) “Watson” like systems like Isabel Healthcare are on the horizon for helping people make informed decisions, and will be disruptive. It strikes me that some combination of FitBit data collection with the power to harness and store data like Google, and natural language processing like Siri, will let some startup really advance this space. Same thing for other fields that have yet to be disrupted by SuperCrunching.
4) SuperCrunching tends to beat expert opinions. Much as chess engines have recently surpassed chess experts in ELO ratings, SuperCrunching will probably surpass human experts on many things.
5) Can entrepreneurs capitalize on the intersection of SuperCrunching feeding into Choice Architecture (Nudge)? Like Medicare Part D being in a recursively improving Choice Architecture.
"The reason for this low participation rate and the many problems associated with Medicare Part D can be summed up in a single phrase: The plan is far too complicated." - Investopedia
6) The ability to crunch down to a specific number is compelling and easy for people to grasp. Wine scores and Sabermetrics (cited in the book), but also Apgar Scores and Klout come to mind for me.
7) Saying a politician is favored 52% to 48% with a margin of error of 3% is less compelling and understandable than just saying Politician A has a x% chance of winning, which you can get from the same data.Although even probabilities are not that easy for people to grasp (need reference from Kahneman's Thinking Fast and Slow")
8) Bayesian probablities (also mentioned in Drunkards Walk) are really important to understand yet poorly understood. It’s about how one piece of data can give you information about how other probabilities change. The example cited was how doctors, even with test info on Down’s Syndrome likelihoods in babies in the womb, don’t know how to combine info to come up with a “probability that your baby has Down Syndrome”
The math is not hard (multiplication and division) but the intuition eludes most people.
9) “Intuition is an input into SuperCrunching” and “Intuition is still important” both struck me as true. Much as Einstein used intuition to come up with mental leaps. Then, as an example, called on astronomers to test that gravity bends light. Intuition remains important even as SuperCrunching rises.
From Walter Isaacson's book "Einstein: His Life and Universe"
One day during the 1930s, Einstein invited saint-John Perse to Princeton to find out how the poet worked. "How does the idea of a poem come?" Einstein asked. The poet spoke of the role played by intution and imagination. "It's the same for a man of science," Einstein responded with delight. "It is sudden illumination, almost a rapture. Later, to be sure, intelligence analyzes and experiments confirm or invalidate the intuition. But initially there is a great forward leap of the imagination."
10) I have a hypothesis that SuperCrunching will come down in price and be outsourced, much as Amazon Web Services has driven down the price/complexity of web hosting. I think smart entrepreneurs will be able to take advantage of this to move more deftly through the fog of entrepreneurship - what’s right/wrong. I think Bob Gilbreath’s Minimum Viable Concept idea is along these lines.
Inc Magazine "[The Minimum Viable Concept is]a test that helps identify promising start-ups even before they create that minimum viable product. This saves investors and companies time and money."
--- Let me know your thoughts: firstname.lastname@example.org and @howierhee