The Art Of Business, Where Science And Business Depart.

Today’s essay is a long one as well, so in case you’re pressed for time, I’ll give you the TLDR here.

If you want to read the essay just skip this section.

TLDR: Science relies in large part on replicability. Business does not. Copying is a zero-sum game and all participating players will compete profits away until they reach a Nash-equilibrium, where a further lowering of price and thus margins is unsustainable. Innovation is always new and therefore never formulaic. There are certain guidelines successful businesses seem to use but in the end, you’ll still have to go with your gut and make a judgment call according to your own ethical values.

Arguably, replicating findings is one of the best tools we have in our scientific arsenal in order to test the validity of a conclusion.

Which is why the psychology community was not amused when the methodological crisis now known as the Replication crisis started, where it was discovered that many studies weren’t replicable (Nosek et al., 2015).

Even some of the most cited and well-known studies such as the study that power posing makes you act bolder (Carney, Cuddy & Yap, 2010).

Which, incidentally, is why it’s so important to not just reference research (instead of parroting some YouTube influencer or blogger) but to also carefully analyze the literature and its limitations yourself.*

This doesn’t guarantee perfection, but it does decrease obvious and preventable mistakes.

JP Ioannidis pointed out serious problems with published research in ‘’Why most published research findings are false.’’

‘’ Simulations show that for most study designs and settings, it is more likely for a research claim to be false than true. Moreover, for many current scientific fields, claimed research findings may often be simply accurate measures of the prevailing bias.’’ (Ioannidis, 2005).

Methodological issues aside, replication matters.

Well, that’s all fine and dandy weren’t it for the pesky fact that our field relies on innovation.

Even if you extract all the knowledge from the brains of Kevin Systrom and Mike Krieger and build Instagram, you’d get the completely opposite result because Instagram already exists.

Even Kevin and Mike can’t replicate it because the variables have changed.

The entire environment and current social climate (regarding privacy, advertising, users as the product and social media addiction) are different, and even if it weren’t, no user is gonna switch to an IG clone when IG already exists.

So what makes business so complex is that it’s both art and science.

There are some guidelines that most successful companies seem to follow, however, you can’t formulate a step-by-step plan for the reasons laid out above.

Peter Thiel was one of the first to point out this distinction between companies that go from 0 to 1, innovation, and companies that go from 1 to n (copying), in his book Zero to One.

Companies that do copy and thus employ a formula they copied from another company open themselves up to fierce competition.

As more players participate in this zero-sum game, the fixed pie of users and profits will get competed away because there are too many mouths to feed with finite resources.

If you copy each other with no (or no meaningful) distinctions than price will be the only thing you can compete on, which is exactly what we see in commodity markets.

Player B will lower prices in order to steal market share from Player A. Player A lowers them still. Player C lowers it below both of them. Player D goes too low and goes bankrupt.

Over time, price will collapse into a Nash-equilibrium (no player can do better by unilaterally changing her strategy).

A point where no player can go below the current price point without going out of business, or above the current price point without attracting too few customers and going out of business as well.

(Unless one player, usually a new one, sufficiently innovates and thus creates a whole new business model. In effect, changing the game.)

This problem is further exacerbated by the fact that we go through traditional education by default.

A system that’s not designed to teach independent thinking but rather to reward obedience and implementing rules.

If you think that’s my opinion vs. facts then watch this to get a better understanding of how our modern education system was formed:

In mathematical set theory, there’s something called the axiom of choice (AoC).

Without making AoC too complex, picture this:

Suppose you have 3 jars: Jar 1, Jar 2, Jar 3.

Each jar contains a piece of paper with a few whole numbers written on it.

Jar 1: 4, 8, 45

Jar 2: 646, 23

Jar 3: 894, 7, 45649, 54, 76

Now I can make a new jar: Jar 4, which contains one element from each of the jars.

Taking one element out of each indexed Jar and using that to create a new Jar (set) with those elements.

Okay… in order to do that we need to build some machine where we can put something in and it spits something out (a function).

Let’s create a function that just ‘’chooses’’ the smallest element in each jar.

So it picks: 4 from Jar 1, 23 from Jar 2 and 7 from Jar 3.

So now we can put a piece of paper in Jar 4 with 4, 23, 7 on it.

We’ve created a choice function ‘’pick smallest element out of each jar’’ that creates a new set (Jar 4) with the chosen elements.

Even if we had infinite jars, this function always works.

Because with natural numbers (whole numbers 1 and up), you can always pick a smallest.

But what if we have an interval in Jar 2 such as (3,5].

The symbol ( means 3 is not included.

This creates a problem because what’s the smallest element we should pick now?

3.1? Well, what about half that 3.05?

Well, what about half of that again 3.25?

No matter how close we get to 3 we can always choose a smaller element so our choice function doesn’t work here.

When you invoke the AoC you claim that even though we don’t know what the choice function is, you’re going to assume there is one.

(That there is a way to pick one element out of each jar and create a new jar with those elements.)

An analogy Bertrand Russell came up with is that if you have an infinite pair of shoes you can create a choice function: Pick the left shoe from every pair.

But if you have an infinite pair of identical socks, there’s no such thing as right or left, so you can’t create a choice function to pick one sock from each pair and therefore have to invoke the AoC.

The AoC is something that can not be proven or disproven, you either invoke it or you don’t.

If you use it, it allows you to do certain kinds of mathematics but also creates paradoxes such as the Banach-Tarski paradox.

(Hard to explain without mathematics but a sketchy way to think about it is that it’s possible to break a sphere into pieces, rotated them and you get your sphere back, rotate them differently and you have two spheres with the same volume as the original.)

And if you don’t invoke the AoC, there’s certain mathematics you can’t do.

Just like not knowing the existence of a choice function and invoking the Axiom of choice, we often can’t know the best way to proceed in business…

under the current market and technology circumstances.

We often can’t prove or disprove our hypotheses, especially when you’re innovating, so reason fails us.

We talked about the limitations of reason and the upside of using a more evolution based approach in the essay series: ‘‘Why Your Business Needs More Weird Ideas.’’

In essence, you’re saying: ‘‘It’s fundamentally impossible to determine what the best way to innovate is, but I’m just gonna assume there is a way.’’

It’s an enigmatic problem so you just have to take a stance and make a decision according to your ethical values.

I want to end with an excerpt from the interview with Andrew Mason, the Founder and early CEO of Groupon.

The Quick Rise and Even Faster Fall of Groupon, Through the Eyes of Its CEO.

There was one other part of the letter that says your biggest regrets are the moments that you let a lack of data override your intuition of what’s best for the customers. What did you mean by that?

Groupon started out with these really tight principles about how the site was going to work, really being pro-customer. And as we expanded people in the company would say, “Hey, why don’t we try running two deals a day?” “Why don’t we start sending two emails a day?” And I’d think, That sounds awful. Who wants to get two emails every single day from a company. And they’d be like, Sure, it sounds awful to you. But we’re a data-driven company, so why don’t we let the data decide? Why don’t we do a test? And we’d do a test, and it would show that maybe people would unsubscribe at a slightly higher rate, but the increase in purchasing would more than make up for it. You’d get in a situation where it doesn’t feel right, but it does seem like a rational decision.

The problem was when you’re in hypergrowth like this, you don’t have time to see what is going to happen to the data in the long term. The churn would catch up with you. People would unsubscribe at higher rates, and then, before you know it, the service has just turned into something — if you look at Groupon now, it’s just this vestige of what it once was. There’s no real copywriting. It’s a marketplace of coupons. It’s still a service that a lot of people get a lot of value out of, but it doesn’t have the spirit it once did.

There are certain things you have to be religious about in the company. That’s what I’ve taken away from that: There are some things where you have to say, “I’m sorry. I’m not going to look at the data on that. This is just what we’re going to do. We know that it’s right, and there’s nothing that’s going to shake us from that.” (Blumberg, 2018).


*The fact that almost everyone is lazy and assume others have done their homework vs. doing the hard work of checking it themselves is a cognitive error known as diffusion of responsibility. ‘’The individual assumes that others either are responsible for taking action or have already done so.’’


Nosek, BA et al. (2015). Estimating the reproducibility of psychological science. Science, 349(6251). doi: 10.1126/science.aac4716. Retrieved from

Carney, D. R., Cuddy, A. J. C., & Yap, A. J. (2010). Power Posing: Brief Nonverbal Displays Affect Neuroendocrine Levels and Risk Tolerance. Psychological Science, 21(10), 1363–1368.

Ioannidis JP. Why most published research findings are false. PLoS Med. 2005;2(8):e124. doi:10.1371/journal.pmed.0020124

Blumberg, A. (2018). The Quick Rise and Even Faster Fall of Groupon, Through the Eyes of Its CEO. Retrieved from

Originally published at on June 18, 2019.



Make a living off your expertise:

Get the Medium app

A button that says 'Download on the App Store', and if clicked it will lead you to the iOS App store
A button that says 'Get it on, Google Play', and if clicked it will lead you to the Google Play store