It is very tempting to build something that users say they want. It’s almost as if they are giving you the answers to the test! Here’s the problem with that; there is (often) a huge divide between what people do vs what they say. Okay, so what’s the worst that can happen if you design for what people say vs what they do? 

Ask Walmart. 

Back in 2009, Walmart executives were looking for new ways to increase profits by experimenting with their store layout. One huge difference with their rival, Target, was how packed Walmart’s stores were. So they asked their customers, “would you like Walmart aisles to be less cluttered?” The majority of responses was yes. 

This led to a major multi million dollar retrofit of their stores, reducing shelf space by 15%. 

The result? 

Almost 2 billion in lost revenue occurred over the course of 8 years before returning to their original store layout. 

So what went wrong? 

There were two glaring flaws in their approach.

1. As mentioned before, they were approaching the survey assuming that what people say is the same as what people want. 
2. Those who took the survey were given the option of a yes or no answer, giving no context for their response. Don’t try to validate your hypothesis with a yes or no question.

These two factors became a recipe for disaster for Walmart. Surveys are useful if you know how to use them  however they will always leave out the most important facet of the data, the why. Talking to your users and mining the conversations for insightful gems can really help your decision making process.