Spherical Cows and Ruling ThumbsTyler WeeksBlockedUnblockFollowFollowingJun 20If you’re hoping to move beyond good PowerPoint slides and start influencing here’s the takeaway: a good rule-of-thumb is that an analysis isn’t done until you have a rule-of-thumb.
There’s a joke us nerds like to tell about ourselves that starts with a dairy farmer looking to improve his milk production.
He goes to a physicist for help who stares at her chalkboard for a second and asks, “Is it reasonable to assume your cows are spherical and uniformly filled with milk?” The joke is funny because this handwavy approach to problem-solving is exactly how a real physicist in real life would probably go about it.
They might think, “Well that’s a hard problem.
That’s too much work.
This other problem is a lot easier and if I squint it’s basically the same.
” Sorry to burst your bubble but science isn’t exact.
In fact, the best researchers have a knack for finding the spherical cows in every problem.
Spot the cow.
gifFrom Wikimedia Commons, the free media repositorycommons.
orgEnrico Fermi won a Nobel Prize for his work on nuclear reactions but he was famous among his colleagues for his powers of estimation.
The story about his visit to the Trinity nuclear test site became one for the record books.
Those tests were expensive and the goal of the experiments was to maximize the yield of the reaction.
Wanting to get a quick sense for the success of the test he dropped small scraps of paper as the atomic wind passed.
He looked at how far the wind carried the paper and used a few back-of-the-napkin calculations to make a quick guess.
Remarkably, when the labor-intensive calculation was completed weeks later he was within spitting distance of the final result.
The beauty of his approach is that he was able to start making decisions and begin planning for the next test immediately.
As data scientists and analysts working in a business environment we often get requests similar to the dairy farmer’s question.
For example, we may be asked to calculate the return on investment (ROI) of a policy or purchase decision.
A real-world ROI calculation can be as complex as calculating the yield of a nuclear test with labor costs, depreciation, satisfaction scores, quality, and who-knows-what-else coming into play.
At a fast-paced company you don’t have weeks to explore hundreds of inputs, collecting decimal points like they were Pokemon.
You need answers quickly and the best place to start is with a few basic assumptions, your spherical cows.
The goal of this blog post isn’t to teach you how to make good assumptions but this textbook will if you are interested.
What I do want to do is to give you some storytelling advice; don’t throw away those cows.
com/1445/Your customers are trying to make decisions in a world where failing to make a decision may be the costliest decision of all.
What they want from you is not answers; they want help making decisions.
The spherical cows you’ve collected not only help you deliver a recommendation quickly but they themselves are a valuable outcome of your analysis.
With a little sprucing up a spherical cow can become a rule-of-thumb.
Numbers are meaningless without context, which is why storytelling is such an essential element in the data science toolbelt.
A rule-of-thumb can be a powerful storytelling device to give your results some context.
For example, knowing that spending more on social media ads did in fact yield higher than expected sales for the quarter is helpful.
The decision was a good one.
However, it may be even more important to know that your company gains roughly ten followers for every thousand dollars you spend on social media (btw: totally bogus numbers so don’t use that).
With this rule-of-thumb you’re giving your stakeholders something better than an assessment of a past decision; you’re giving them a tool to make the next decision.
Your stakeholders are more likely to see you as an expert if you set them up to look like experts.
The rules-of-thumb you develop are important parts of the story that they can use to influence their peers and leaders.
If you’re hoping to move beyond good PowerPoint slides and start influencing here’s the takeaway: a good rule-of-thumb is that an analysis isn’t done until you have a rule-of-thumb.
.. More details