As a physicist-turned-entrepreneur I think a lot about about how the techniques that I've learned in science can help me to build a successful start-up. In that regard, one of the most useful rules I've learned as a scientist is probably this one:

Make it work on paper first.

What does it mean? Just like a successful entrepreneur, a good scientist continuously has lots of ideas for new experiments. The problem is that it's usually impossible to turn all these ideas into reality due to time and money constraints, hence a process is needed to weed out bad ideas as soon as possible. For this, scientists often use a method that is often attributed to Enrico Fermi but which was probably used by countless scientists before him. It works like this:

  • Determine the quantity you want to estimate
  • Identify the most relevant variables that determine this quantity
  • Estimate the rough value or order of magnitude of each variable
  • Calculate the searched quantity using these estimates.
A famous example of this method is Fermi's estimation of the energy released in the Trinity nuclear weapons test in 1945, where he used an "order-of-magnitude" experiment to estimate the yield of the nuclear explosion:
About 40 seconds after the explosion the air blast reached me. I tried to estimate its strength by dropping from about six feet small pieces of paper before, during and after the passage of the blast wave. Since, at the time, there was no wind, I could observe very distinctly and actually measure the displacement of the pieces of paper that were in the process of falling while the blast was passing. The shift was about 2 1/2 meters, which, at the time, I estimated to correspond to the blast that would be produced by ten thousand tons of T.N.T.* (from Richard Rhodes: The Making of the Atomic Bomb, p. 674)
OK, estimating the yield of nuclear weapons is not so important to startup founders these days, so let's look at a more relevant example. Let's say we have an idea for an online community about books, where people can share what they read and find interesting books to check out. And let's assume we plan to refinance that through affiliate links (e.g. to Amazon). How many users do we need to make 10.000 $ / month with this business model? Let's check!

Quantity we're looking for:
Number of users required to earn 10.000 $ / month through affiliate marketing of books

Relevant variables:

  • Average number of books a given user buys in a month: 2 - 4
  • Average price of a book: 10 - 15 $
  • Average commission earned through the affiliate network: 5 % of wholesale price
  • The percentage of book purchases a given user makes through our website (relative to all his book purchases): 30 - 50 %

Of course, the numbers given here are just rough estimates. OK, so what does this give? On the lower end, if the average price of a book is 10 $, the commission 5 %, the probability that a user will buy through our site 30 %, and the number of books purchased in a given month equal to 2, we will earn on average 30 Cents per user and month. On the higher end, with an average of 20 $ / book, 6 books per month, and 60 % purchase share, we will make 1,50 $ per user and month. This means that we will need roughly between 6.700 and 33.000 active monthly users to make 10.000 $ per month, depending on how engaged our users are with our site and how many books they buy on average.

Looks really simple? It is! Unfortunately, many people tend to omit this step when thinking about their experiment (or their start-up), so they end up spending months or years on an idea that can't even work out on paper. So, next time you think about a new experiment or a start-up idea, make it work on paper first before starting to build the real thing.


* the real yield of the nuclear explosion was later determined to be around 18.6 kT, so Fermi's first estimate was remarkably close to the mark