When you have a startup idea, the first instinct is to pitch it to friends and colleagues. This can be somewhat helpful, because they may provide insight that you hadn’t thought of yet, or quickly validate some of your concerns based on their questions. They may be familiar with a competitor worth looking at, or know people in the space that can help. Or they may be able to brainstorm with you and “sniff test” your ideas, which can be revealing about your thought process and where you need to focus. But unless they’re potential customers, it doesn’t really count as quality feedback.

Hopefully most entrepreneurs realize they need more feedback than what they can get from friends and colleagues. Unfortunately, many will want to throw up an online survey and drive traffic to it, hoping to get respondents and quality information. This isn’t where you should start.

The problem at this stage is that you don’t even know what questions to ask. And a good survey assumes you do. It assumes you know what you want to ask and that you’re nearing the stage (or you’re at the stage) where quantitative survey data is helpful for validating or invalidating your assumptions. Before you get to doing a survey — which can be helpful — you need to meet people in-person.

I realize this is tough. You may not have experience interviewing people and interpreting feedback. You may not be sure how to find people to interview. But it’s an incredibly valuable step. It’s the most effective way to get honest, open-ended feedback from people (who you think are in your target market). And it will also increase your own personal confidence and communication skills. It’s almost guaranteed that the original hypotheses and assumptions you had going into the in-person interviews won’t survive, but it’s almost guaranteed that you’ll tease out interesting ideas, trends and new opportunities. You won’t get that kind of rich qualitative data from surveys.

Set a goal to speak with at least 20 people.

Find them by any means necessary. If you have to pay them, do it. They’ll still be honest with you. People love to talk about their problems. Don’t be shy.

By the tenth person you’ll see patterns in people’s answers, and you’ll start adapting your questions accordingly. You don’t need to speak with hundreds of people to get what you need – which is ultimately validation (or invalidation) of your idea. Refine the qualitative data from in-person interviews into more pointed, quantitative survey questions and then try and attract a larger audience to take a survey. The data collection from the survey can further support the findings of in-person interviews. But the process has to start with in-person interviews.