That’s Not a Hypothesis

Product growth is science. Without the scientific method, we’d just be “growth hacking” — trying random things from loud blogs without learning about our users.

The cornerstone of the scientific method is forming a hypothesis. That’s a nice sentence that leads everyone to nod their head, yet a stunningly high percentage of brilliant product people don’t write real hypotheses. Instead, they write predictions, or solutions, or don’t bother with hypotheses at all.

A good hypothesis is a statement about what you believe to be true today. It is not what you think will happen when you try X. It contains neither the words “If” nor “Then.” In fact, it has nothing to do with what you’re about to try — it’s all about your users.

Why be pedantic about this? Because hypotheses are the key to learning. Product growth doesn’t happen from a few cool tricks. Product growth comes from fumbling around in the dark, trying a lot of things, and improving our aim over the course of months and years. In other words, this is a long game that is ultimately about learning. Clear learnings come only from clear hypotheses.

Let’s Talk About Science

Dr. Brand agrees, this would be a good time to talk about science.

Science is about building a model in your mind about whatever it is you’re fanatically curious about.

That’s a cell, an electron, a species — these are things you can’t look at all the time from all directions— you have to make a model.

In product or growth, that’s your users. How do they think? What do they care about? How do they behave? What don’t they know?

Over time, this model gets better and better as we stumble around, try experiments, and observe more and more. We clean out wrong assumptions, and add texture to facts that appear true.

Thus, a Hypothesis Is…

A hypothesis is one aspect of that model in your head. It is a something you believe to be true. Maybe you believe it so much that you call it a fact, because it’s staring you in the face. Maybe it’s something you believe is such a logical extension of what you know for sure, that it must be true.

The point is, it’s not about your experiment or what you think is going to happen. It’s about what you believe is true about the thing itself:

  • We believe that plants need sunlight to live
  • We believe that electrons have a negative charge
  • We believe that white blood cells defend the body against illness
  • We believe that patreon creators are highly motivated to use our platform, but experience writers’ block when building their pages

Even if they’re painfully obvious and true and have a ton of evidence behind them — they’re still hypotheses. They’re still things we believe.

Hypotheses should be falsifiable — that means there’s something out there you can practically observe that would lead you to reconsider the hypothesis. Saying that the Earth was created 6,000 years ago, but made to look much older, is not falsifiable.

This is, in fact, a well-stated hypothesis, and one that’s falsifiable.

Where Do Hypotheses Come From?

Hypotheses have reasons behind them. In growth, that could be past experiments. It could be user research. It could be a conversation we had with sales. It could be a particular metric.

Or you could know SO little about something, that it’s a wild-ass guess of a starting point. That’s cool too, as long as you are conscious about it. As Richard Feynman, Nobel Prize-winning physicist explains it:

It is not unscientific to take a guess, although many people who are not in science believe that it is.
Feynman is my hero. Start here.

Good hypotheses cite sources: Why do you believe this to be true?

  • We believe that plants need sunlight to live because plants do not seem to grow in places where there is no light
  • We believe that electrons have a negative charge because when they shot through a magnetic field they curved to the right instead of going straight.
  • We believe that white blood cells defend the body against illness because ill patients appear to have a higher count in their body compared to healthy patients.
  • We believe that Patreon creators are highly motivated to use our platform, but experience writers’ block when building their pages because when the customer success team helps a creator launch, most of the time is spent helping them find relevant examples.

A Prediction Is…

Now that we have a clearly articulated hypothesis in hand, a prediction is what you believe will happen if your hypothesis is true.

This is what many smart people mistake to be a hypothesis.

  • If I put my plant in the dark, then it will die
  • If I receive an infection, then my white blood cell count will go up
  • If we put concrete examples in the Patreon onboarding, then we will see a rise in successful creators

There can be many predictions from one hypothesis. One belief can lead you to try many things.

This is actually a prediction, not a hypothesis!

Clear Hypotheses Lead to Clear Learning

If all you have is “If I put my plant in the dark it will die” (a prediction) and that turns out to be false, then all you learned is… that sentence was wrong.

Suzy knows the importance of clearly stating a hypothesis.

However, if you started with “Plants need sunlight to live” and your particular plant survived in the dark (!!!), you now know that statement is not exactly true. Maybe there’s exceptions or nuances. Maybe there’s a new experiment to try. Maybe it’s time to articulate a new, more subtle hypothesis, or a crazy new one.

If it turns out your prediction was wrong, it calls your hypothesis into question. Maybe it turns out to be completely false (happens to me a lot). Maybe only parts of it are incorrect, and you simply need to add more subtlety to it (also happens a lot).

These also happen to be the most exciting moments in science. (Surprises are good, they lead to Nobel Prizes).

Your mental model should get better and more refined over time. When things get really confusing, you might need to completely overhaul it.

In other words, without a hypothesis, we didn’t learn anything.

Side note: What About Problem Statements?

As growth and product people, we’re paid to solve problems. They’re usually simple to articulate: “Users are not doing [thing we want them to do].”

  • “We’d like customers to retain longer.”
  • “We’d like users to engage more.”
  • “We’d like visitors to convert at higher rates.”

Problems are our area of focus, our initiative, our objectives. To solve them, we start by figuring out why something we want to happen is, well, not happening. Hypotheses explain why the problem is happening.

An overview of how the Patreon growth team thought about new creator onboarding. Btw, this was done on Freemind, my favorite piece of software ever.

When we structure our thoughts like this, wonderful things happen. We focus on simple solutions first. We keep scope small. We know exactly why we are doing what we’re doing. Ideas are straightforward and simple. Scientific process is a beautiful thing.

Your Checklist for Writing a Good Hypothesis

I’ll leave you with a quick rule of thumb for ensuring you’re articulating good a hypothesis:

  • It starts with the words “we believe”
  • It probably has the word “because”
  • It doesn’t contain “if” or “then” (that’s a prediction)

That’s it. Now let’s go learn!