The short answer to this question is easy: You don’t know if an idea is worth pursuing. You simply can’t know for sure that your nascent product or service is bound for success. If you could know this with certainty, then you wouldn’t bother going through the trouble of spinning up a factory, writing your code, or patenting your widget. Just retire to your basement and day-trade stocks; clearly you can see the future!

While the latter response is a bit tongue-in-cheek, there’s a lot of truth in it. Of course it’s important to do research so that your best guesses are more than guesses. Likewise, it’s critical to ask your potential customers what they want and (perhaps more importantly) to anticipate better than they can what they might want in the future. But it’s very hard to predict what the market will embrace and what it will not. Enormous companies with enormous budgets frequently launch products and services that fall flat when shown the light of day. If they can’t predict how things will turn out, what chance does someone in a garage have of doing better?

See also: The Easiest Way to Know You’ve Got a Business Idea Worth the Investment

Here’s something I find valuable: Rather than think about an idea that is worth pursuing, think of an experiment that is worth doing. While the differences between these two options may seem negligible, the latter way of framing your next course of action helps to reinforce a mindset of rapid prototyping, quick market entry, intense data collection, and swift analysis. Perhaps more importantly, it also helps to keep you from becoming wedded to an idea in an unhealthy way. We look at our ideas like children we can’t let go of—children that the market might reject. That hurts.

Experiments, on the other hand, are temporary efforts that tell us something about what might (or might not) work. There is no real failure of an experiment. If structured right, an experiment tells us whether the underlying hypothesis tests true or false. A “true” indication tells us the path we are on might work. A “false” indication tells us that we need to head in another direction or tweak the path we are on. It’s much harder to be flippant about an idea that didn’t work. In that case, many might say to themselves one of two things. The first response might be: “My idea didn’t work. I’m not good at this.” The other is perhaps more dangerous: “The market doesn’t get it. I’m going to keep spending time and money to convince them they are wrong.”

When I was a student getting my MBA at Harvard, I had a professor who said that this “arrogance of the inventor” is one of the most common reasons an idea dies before it ever really has a chance. If you keep pushing on the customers your widget with feature A when what they really want is your core widget but with feature B, you will lose. You will lose because you didn’t iterate after the pain of rejection or you will lose because somebody else in a garage down the street took your idea and augmented it with feature B to gain the share you had within your grasp. Thinking of your path forward as embodied in experiments rather than in ideas will protect against this arrogance and increase the likelihood that you quickly get to a place where your ingenuity meets market need.