Joyeeta Das | Founder & CEO, Gyana Limited

Fermi question - passion in a startup: what connection?

Do you know how many cup-cakes will be sold in London in the month of January? If you knew even to some extent - how would that change your inventory forecast?

Enrico Fermi could calculate the size and impact of the Atomic bomb by using the distance the papers that flew from his hand, travelled to. He said "it is 10 kilotons" and currently scientists estimate the answer close to 17 kilotons.

Thus, by using what he knew, he created a way of knowing what he did not know- remarkably accurately. This is the fermi question, Fermi Law or Fermi estimation method. It helps to breakdown complex problems into simple problems and solve them step by step.

A classic use of the Fermi Question method would be “how many piano tuners are in Chicago” or “how many cars are there in Baltimore”?

Today many “consultant firm job applicants” try variations of this problem in their “case practice”. (They may or may not know this!)

Basically, you use what you know to arrive at a “back of envelope” calculation that is 10 orders of magnitude closer to the correct number - why 10/20 orders are useful at all? Because otherwise in your enthusiasm and oversight you can end up with beyond 20 orders of magnitude of error. Many scientists use this to guess-estimate the order of magnitude of their results before they arrive at it. Many experts exist in this area and there is a whole universe of Fermi Question competitions and so-on.

The fine art of guessing is, thus, a science.

You start with what you know, use the variations to upper and lower bound it, and use common sense to validate. The over and under estimates cancel each other out.

Say for the piano problem - you go - what is the population of Chicago - what part of that population may use pianos- what percentage need servicing - how many can a piano tuner service in a day and hence what is the demand etc.

Then use common sense to make a few adjustments here and there.

This is also the reason why entrepreneurs who start off from a personal experience tend to be more successful and passionate. When they personally go through an experience and make a business out of it, they tend to have a good understanding of the “first estimate” or seed estimate. The most efficient big data engine, their mind - tends to have picked some future numbers of a global market that the conscious mind may not know about even. The better the seed estimate, the better the entire process of the guess. They say the future of the Universe may have been determined by the first 10 seconds of the Big Bang! Hence the seed estimate is extremely important in how well this process goes. (And there you were, thinking passion is purely an emotion fuelling the business)

Entrepreneurs who start off from a personal experience tend to be more successful and passionate

The entire art and science of business is a lot of smart guessing - I use Fermi a lot and this validates and makes my business decisions stronger in addition to bolstering my confidence.

When we start purely from a “non - passion” mode, such as “this is a market gap I must leverage even though I just do not care” or “I have no understanding or experience of this area but I will start anyway” - chances are, it may end up with a gross misunderstanding of the future of the business in terms of market sizing and thus funding, hiring, selling the product (basically everything). This does not mean exceptions do not exist - I am sure some have still made it despite no prior personal connection to what they are solving. I bet they are business genius! (I am not).

However, like all processes you may go wrong with Fermi in 3-6 major ways. Choosing a non-linear problem is classic amongst those, modelling the problem incorrectly is another.

Mathematically - my write up above is not perfect. I must admit I have omitted the scientific basis and the error propagation techniques and simplified the method a lot. However, even if you are not a scientist - by practicing basic Fermi equations you can make much better business decisions, you do not have to be a mathematical expert.

  1. Google up - start with a few basic Fermi practice questions
  2. Apply to your own business for decisions in the past
  3. Estimate casually why the result is/was divergent from your number
  4. Use in future.

That is all! No logarithmic calculations- no derivations - nothing.
However, if in future you do end up trying this and are baffled, hit me up. I love to help.

Happy scientific guess-estimating in business!