Decision over Decimals: my Synthesis

It’s better to be roughly right than precisely wrong John Maynard Keynes

After the content's usefulness in a book, my main criterion is efficiency: if you can reduce its 200+ pages to a half page blog or, even worse, a single sentence, then reading it is a waste of time. In contrast, ‘Decision over Decimals’ is worth your time: it is informative and each chapter makes distinct and interesting points. It does have a central theme (Quantitative Intuition, trademarked), but generates at least a dozen key recommendations based on the rich blend of practical and academic know-how.

So here is the key message: strike the balance between intuition and information, which are supposed to energize each other (just as in Amazon’s work-life balance, not ‘harmony’). Decisions OVER decimals means intuition is key, because it is paramount in the first and last step – the most important of the 5 steps of decision making:

1.????Define the problem I

2.????Data discovery Q

3.????Data analysis Q

4.????Insights and delivery I

5.????Implementation I

Each of us will like specific chapters more depending on experience and needs. My personal favorites include this dozen:

1)???The right questions to ask:

Help me understand?

Have you considered?

What surprised you most?

These help the decision maker focus the analyst on surprising findings and outliers, which may identify problems with analysis or winning solutions

2)???The ‘I wish I knew’ (IWIKTM) framework organizes questions along the dimensions of need to know for the decision vs currently know, with the advice to spend more time on the known unknowns than on the unknown unknowns

3)???Working backwards by creating a decision tree, writing the blueprint of the report before you have the data, and reverse engineer the data and analysis map

4)???Plotting the data before analysis, as illustrated by the Anscombe’s Quartet, where all 4 datasets have the same average and variance of X and Y, and the variables have the same correlation and thus explained variance with the exact same regression line:

5)???Focus on verifying the 20% of the numbers that matter, and you can do so with guesstimates to quickly evaluate whether they are in the right ballpark. This suffices for most decisions – it does not make sense to be super precise about a highly uncertain future projection. Likewise, we don’t demand the engineer knows exactly how long a project will take – we simply want to know whether it’s hours, days or weeks. Instead of precise garment measures, T-shirt sizes (small, medium, large, extra large) suffice.

6)???Synthesize and put the recommendation first, then follow with the data-supported explanation

7)???Measure trust in the data and the person who is the source of the data, especially when the decision is a one-way door (hard to reverse)

8)???Build your story arc and anticipate questions by different stakeholders and people who play different roles

9)???Make your decision like a triathlete: swim without drowning in data, bike with precise analysis, run with focus of mind and energy to battle your team’s exhaustion

10) Create the case for the decision, frame the outcome and seek consent, not consensus

11) Tack your big change into smaller changes, requiring fewer people for faster decisions

12) Hire for precision questioning, contextual analysis and synthesizing skills

With Amazon Scholar Oded Netzer as coauthor, you will not be surprised to see Amazon coming back several times: in working backwards with a PRFAQ (Press Release and Frequently Asked Questions), Two-way Door decisions and the Leadership Principles. However, I saw our shared Amazon adventure at several other times:

1)???Listen before you talk (‘Respond, don’t React. Embrace Silence): our first shared manager Saeed Bagheri, basically epitomizes this insight. We once had a bet we could make him react with disappointment. We told him at the start of the meeting that we had no new insight so could give all the time back, and his lack of reaction still brings a smile to my face. Saeed embraced silence until we confessed to the prank.

2)???Support your reports and ask the right questions: Amazon managers such as Sherry Marcus and Dave Zimmer excel in creating an inquisitive culture by asking the right questions and supporting employees who take initiative. The ‘Day 1’ feel of the place means mechanisms to avoid the ‘Not Invented Here’ Syndrome the book warns about.

Happy reading and implementing this wonderful book in your business!