The curse of the validation mindset

And the other muscle we never train for

Somewhere at a university, a workshop room awaits a new cohort of founders. Post-its and pens are scattered on wheeled tables, with a bookshelf holding dozens of startup books forming the backdrop. Slowly, the ambitious entrepreneurs trickle in, taking spots at their workstations for the day.

The instructor opens with a keynote stipulating that most startups fail, and therefore jumps to the importance of validation. Your ideas are a set of hypotheses, he says. You need to isolate them, design an experiment, run it, and presto: you can validate your idea. The founders follow the instructions and begin jotting down their hypotheses on the sticky notes.

To most of you, the above might seem normal, even sensible. But this is exactly how people are cursed with the validation mindset. And that curse is more likely to kill momentum than enhance it, because it assumes hypothesis generating is the easy part. It isn’t, and here’s why.

Validation limits thinking

The hypothesis validation mindset, popularised in Eric Ries’ Lean Startup, echoes scientific principles. Formulate a hypothesis, design an experiment, run the experiment, and validate or reject the hypothesis. The problem, however, lies in the reasoning of hypothesis testing itself.

Validation suggests a yes-or-no question. That means you should be able to formulate your hypothesis in such a way that you can test it. However, based on my experience, asking a founder to formulate a hypothesis in the early stage nearly always leads to the formulation of an uninteresting or poor hypothesis: “Do people hate waiting for the bus?”. What people do is take their startup idea and transpose it to a yes/no question, for which we often have the answer ready.

This is because the validation mindset forces entrepreneurs to elicit inductive or deductive behaviour. Inductive means to go from available data to a rule: I’ve seen the street wet after it rained so many times, so I understand that rain causes wet streets.

Deductive reasoning means going from an available theory to new knowledge. I walk out my front door, see a wet street, and based on my generated theory, I know it has rained. I didn’t generate a new dataset; I could explain a singular observation through existing theories.

Startups need new thinking

But what if my neighbour just washed his car? Am I safe to refute that theory because I have something in place? Or have the neighbour’s kids held a water balloon fight? All of a sudden, there are multiple hypotheses for a single observation.

Notice what I just did? I reasoned backwards from one observation to generate several plausible explanations. That’s the third kind of reasoning: abduction (coined by Peirce). The big difference is that induction and deduction stay within the evidence available, whereas abduction generates new ideas.

This is exactly how startups work. A founder observes something inefficient and wonders why it can’t be improved by adding a new product (a hypothesis). But the true nature of that surprising observation isn’t fully available to the founder. He or she has seen it a few times and can say there’s something funky with waiting for the bus. Is it the wait time? Or is it the uncertainty of it potentially not showing up? Or is it the lack of shelter?

The muscle you haven’t built yet

The problem, however, is that generating abductive explanations for a weird situation is incredibly tricky. For the wet street, we can imagine dozens of explanations because we’ve roamed this planet for at least two or three decades. That builds a muscle to explain stuff, as we understand how rain and water work, for instance.

However, for your new startup idea, you very likely don’t have that track record yet. Therefore, the number of plausible hypotheses you can generate is either limited or of poor quality.

There are exceptions, though, and that is when I meet founders who start a business after working in the industry for ten years. They know what is up and can formulate those assumptions. However, they don’t need to. For them, it was never a hypothesis to begin with.

Why do we do this?

In most Western societies, we love to educate people in analytical thinking skills. Universities are the prime example. In training scientists, the validation aspect takes centre stage. You learn to write papers, and most attention goes to the justification of your findings through rigorous methodologies and data. So hordes of people have trained the justification muscle, rather than on how to generate the best ideas.

And these people then arrive at a startup, or give workshops on how to do startups. They use their best-trained muscle, read the Lean Startup, and love how it resonates. The validation mindset is just an echo of this scientific justification mindset. That explains the appeal: people feel at home in the seemingly safe environment of justification. But the feeling of ease is misleading, because the early stage of a startup is all but certain. By applying that mindset, people aim to tame uncertainty with rigorous induction and deduction; however, it is best treated through abduction. Nobody trained them in that.

That’s why I see so many (university) incubator and accelerator managers loving the validation approach. These people are not selling snake oil; they were just teaching with the muscle they had been trained in.

Startups are not scientific theories

That validation muscle is not equipped for startups. Scientific theories aim to describe the regularities of the world, whereas founders aim to change something about that very world. That transformative process doesn’t happen by itself if you have twelve hypotheses validated. You need to engage with the world.

I’m not advocating for ignoring evidence. However, signals in a startup journey are transient. The fact that you found four potential pilot users in three weeks might sound amazing. However, if that’s the only positive signal you generated after twenty weeks, I raise a red flag.

Evidence evaporates. Markets might change, and your momentum with a customer might drop. Your startup is in flux; it changes continuously. My to-be-published research finds that all your building blocks, including customer segment, value proposition, marketing channel, messaging, and positioning, continuously change form. Assuming you can validate something like a theory of gravity and then move on is a mismatch with venture reality.

What to do instead

Therefore, if I work with founders, I never obsess over the hypothesis. Sure, write down potential market segments; do that for fifteen minutes. But then spend 90% of your time selling to twenty of each segment.

By actively trying to transform the world, you will find which tiny section of the world allows you to. This trains your abductive muscle in two ways. It generates unexpected and surprising findings. And, instead of betting on one hypothesis, you learn about several at the same time. Therefore, the simple act of picking up the phone and selling beats the rigorous, time-consuming setup required for crafting experiments.

After a bunch of such experiences, your intuition will tell you where to go. The validation mindset’s deepest error is mistaking a generative activity for a confirmatory one.