The question that doesn’t help you.

As a point of practice, I participate in almost every feedback request I receive — interviews, usability sessions, surveys and feedback form — anything. And I’ve noticed that, very often, I am asked to predict what I will do.

There are a ton of ways this question might be asked:

  • Would you use [x]?
  • You would use [x]: [every day], [multiple times a day], or [not at all].
  • On a scale of 1–10, how likely are you to buy [x]?
  • Would you pay [x] every month for this service?

Honestly, I understand the urge to include this question. We all want to get a quick temperature check on an idea and know if our audience will use a product. But a quantification from these questions can’t be trusted because people are notoriously bad at predicting their future behavior and accurately detailing behavior too far in the past. Participants often respond with idealized visions, leaning on aspirations and hopes to imagine what will be.

I have an aspirational self. That person is amazing. He exercises every day and he eats perfectly. That person, unfortunately, is not real. The real Brad will definitely skip exercise to go eat pizza…often…and with excitement.

But when I think and talk about what I will do, I don’t think about how I might skip yoga to sleep in or eat french fries instead of salad. I think about my aspirations, my ideal future.

This means that people may tell you the product is perfect for them and they would buy it, but, looking at purchase history, you might find that they have never bought (or even searched for) a similar product. Instead, trust in actions and behaviors and be skeptical of idealized, aspirational rhetoric.

So, what questions to ask then?

Instead of asking about the future, take a recent time period and ask about concrete actions. For example, if you’re making a product for food delivery, you could ask “how many times did you eat out this week?” or “Describe the meals you made at home last week?”

I’ll note here that relying on the past as an indicator of future behavior is tricky and has its limits. You should only really do so if:

  • The past time period is recent;
  • Most things are equal;
  • Actions are be habit-like (recurring often); and
  • The wanted behavior will happen in the near future

Also, this is always a guesstimate, even if someone has done similiar actions or activities, it doesn’t mean they woudl brunt even a minor change to that routine. I’m a firm believer that people can change, but it’s the exception, not the rule.

In sum, stop asking people to predict the future — it doesn’t help you make better product decisions!


When trying to figure out if someone really will use a feature or product, rely what they’ve done, not what they say they will do.