Paradigm Shift: Drastically Increase The Odds of Success

TLDR: There’s an unconscious and highly dangerous assumption in business that, if one only thinks long enough and executes a rigorous enough analysis, one can derive the perfect answer to any business problem. This is the axiom the McKinsey’s of the world are built upon. To make matters even worse, it’s an extremely powerful delusion to which founders are greatly attracted. However, you cannot know ex-ante, only ex-post. This implies that we must shift our methodology from assuming we’ve got it right, to assuming we’ve got it wrong. Essentially giving ourselves permission to be okay with a high quantity of false negatives; ideas that we discounted that were actually good.

Resisting The Siren’s Song talks about a similarly attractive delusion, namely the importance of a visionary founder in the early days.


In mathematics, you actually can do this. Come up with the right answer if you just think long enough and reason well enough. When you think carefully and you’re equally rigorous, it’ll lead to a proof that confirms the conjecture or shows a contradiction (thereby demonstrating it’s false). [1]

One of the reasons I enjoy working on mathematics is because of this lack of ambiguity. [2]

But this perfect answer, as a result of deep thinking, exists somewhere between hardly ever and never in our entrepreneurial world.

It’s not just that it’s hard to find (which was the essence of the essay How Our Physics Envy Results In False Confidence In Our Organizations), it’s that completely opposing answers to the same problem can be equally successful in the marketplace, e.g. Zara vs. Gucci. That’s what Why Your Business Needs More Weird Ideas was about.

This is especially true once we move beyond the domain of straight-forward optimization problems with clearly definable variables, a precise way of quantification, and a clear ordinality that allows easy ranking from worst to best. [3]

There are simply too many inexcludable variables, the system is highly volatile, and butterfly effects play a large role. That makes it nearly impossible to outthink many business problems. But it’s also just much more cost-effective to simply try solutions in the real world and get quick feedback.

It’s not just the quantity of variables that’s the problem. It’s specifically their inexcludable nature. The linear motion of the system that describes a ball being dropped on the floor also has a great deal of variables. It’s just that nearly all can be excluded while sacrificing near-zero preciseness. I.e. You don’t need quantum mechanics to determine how each particle behaves in order to approximate the system.

Oh, and no… for those smartasses among you. You can’t simply sidestep this issue by asking your customers what they want: Most Market Research Is Horseshit: Why customers couldn’t have told you they wanted a Dyson vacuum.


Let’s assume I’ve done a sufficient job making my case and let the following proposition be true:

It’s impossible or at least cost-ineffective to try to outthink these complex kinds of business problems.

Okay, there are four values we can assign to our solutions/ideas: Positive (good idea), Negative (bad idea), False Positive (seems to be a good idea but actually bad), False Negative (seems to be a bad idea but actually good).

Because of our assumption, we won’t know ex-ante which ideas are Positives or Negatives.

So that leaves us only with False Positives and False Negatives.

That means we’ve got a very important choice to make.

Should our methodology be biased toward False Positives or False Negatives?

Again, allow me to stress that you must choose because if you didn’t, it would contradict our proposition that you can’t know, ex-ante, which ideas are good/bad.

i. If we assume our idea is good until it’s definitively proven to be wrong, we’ll get a high quantity of false positives.

ii. If we assume our idea is bad until it’s definitively proven to be right, we’ll get a relatively high quantity of false negatives. [4]


This is one of the biggest ideas that has come out of startup culture (Blank, 2010). Changing the default from a bias toward false positives to a bias toward false negatives. Unfortunately, it still hasn’t permeated the minds of founders, entrepreneurs, and corporations. I suspect because it is so counterintuitive.

Virtually all founders make the crucial mistake of going with i. ‘assuming their idea is good’.

This is fine if and only if you’re lucky. When you’re not, which is most times, you’ll keep spending resources on a false positive.

This partly explains the extraordinary failure rate in startups as well as the wasted capital in companies like (which raised hundreds of millions).

The only upside is that when you’ve tried everything under the sun to make it succeed, you know with a high degree of certainty that the idea was a Negative.

If you’re a scientist that’s useful. You can publish a paper. Or, in mathematics, you can earn a million dollars by proving the Rieman conjecture or any other Millenium Problem is false.

But entrepreneurs don’t get rewarded for being able to demonstrate something is surely false in a particular moment in time. There’s no upside.

So what’s the antidote?

To do the exact opposite.

Instead of being irrationally passionate about your idea, believing in it, and spending resources on it until you know with a high level of certainty it’s wrong, you flip it. [5]

You assume your solution/idea is probably wrong and treat it as disposable. You test it out in the marketplace and when they don’t respond you tweak it or bin it.

That means that there will be many ideas that’ll get prematurely tossed. But it also means that you’re much more likely to find a Positive because you’ll only move forward when the marketplace responds so you’re not wasting precious resources (time, money) on False Positives.

If you can easily find a paying audience with your rickety MVP and make the unit economics work, there’s a good chance you’re solving a real problem that you can turn into a successful business/product line.

Essentially, we’ve created a methodology that optimizes for identifying Positives (and has False Negatives as a side effect) instead of a methodology that identifies Negatives (and has False Positives as a side effect).


[1] Godel’s Incompleteness Theorem notwithstanding, which showed in 1931 that it’s possible to have a system with a set of non-contradicting axioms and create statements that are unprovable using just those axioms. I.e. It’s possible for something to be true in your system while you can’t prove or disprove it. (Needless to say this turned the world of Mathematics on its head.) Younglings that read The Art Of Business may remember The Axiom Of Choice.

A large part of mathematics is ‘playing’. Trying ideas (sometimes even dumb ones) to see where they lead you. Those can spark some insight that helps you with the problem you were originally trying to solve. In a sense, it appears that even math obeys the idea of embracing failure.

[2] Something that annoys me to an astounding degree is the lack of consensus on even axiomatic things like definitions in psychology and marketing. How can one have productive discourse in such circumstances? Not that this is always the case but, for example, often there’s no precise and uniformly accepted definition for certain heuristics.

[3] Think optimizations in a supply chain or factory for example. It’s precisely this easy numerical expression quality, as well as the straight-forward measurability that creates a kind of quantification bias. A distortion caused by the fact that numbers seem more trustworthy, objective and free of human error. That’s not bad per se, but it’s something to be mindful of. After all, not everything that’s measurable is good and not everything that’s not is bad. I can measure a 10% decrease in the average wait time for a dental practice. But taking steps that reframe waiting into a highly enjoyable experience are harder to quantify yet can offer drastically better returns at a fraction of the cost. Also, Goodhart’s law needs to be taken into account; the tendency of people to start blindly optimizing for a target such that it stops being a good proxy for the thing it was supposed to measure.

[4] The reason I say relatively is that most ideas tend to be bad in the sense that they don’t work out or need pivots (the way Eric Riess defined it). So if you assume everything will fail, you’ll be more likely to be correct than if you assume everything will succeed.

[5] was a False Positive built on the assumption that enough users would order pet toys and other high margin items, which would offset the expensive to ship and low margin bags of pet food and litter. But it turned out their value proposition according to users was having those heavy bags delivered to the front door. Because the margins sucked and shipping was expensive, their unit economics failed to work spending $60M to get $6M. This was completely preventable had they adopted a False Negative approach. But because they decided to sell more than 50% of the company to raise hundreds of millions, taking the time to find product/market fit and pivoting was not an option.


Blank, S. (2010). Steve Blank Why Startups are Agile and Opportunistic — Pivoting the Business Model. Retrieved 15 March 2020, from



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