How to Take a Data-Driven Approach to Customer Research and Overcome Your Biases
When you need to know whether there’s a market and demand for your product or service, and anecdotes from potential customers won’t cut it, taking a data-driven approach can help overcome our natural bias towards self-affirmation. Tom Lowe, senior analyst for WeWork Labs, shares the steps to translate customer interviews and conversations into data points that help you validate the hypotheses around your business.
Why this approach matters
“Customer discovery is about making sure that people want to buy the product you’re selling or proposing to sell,” Lowe says. “The easiest thing to do is sit in your office, read about people’s problems online and seek information that validates what you’re doing. But the only way to truly know if your idea is as great as you think it is have to go out into the real world, talk to people to validate your hypotheses.”
It’s crucial to note that there’s “a distinction between confirming your own biases and seeking the truth in an objective way,” Lowe says. “You’re wasting your time if you’re only looking to validate. And if you find that all of your hypotheses are validated, then you’re either a genius or you’re fooling yourself, and people are far more likely to fool themselves than be geniuses.”
The key is that by following this process, the information you gather goes beyond the notes you wrote or typed during interviews. “If it all lives in notebooks, the ability to transmit that information to stakeholders, like potential investors, is so limited,” Lowe says. Translating the information into data points allows you to “communicate with charts and real numbers, which are much more powerful than anecdotes.”
Step 1 - Create your hypotheses and conduct your interviews
Start by making a list of “all of the hypotheses and assumptions underlying your business model—all of the things that need to be true for your business to be a success.” Lowe says.
Write them down and make them as specific as possible—it’s difficult to validate overly broad or vague statements. “If you have big, broad statements, break them down into theories you’re able to test with facts and numbers,” Lowe says. If your hypothesis is “there’s a great market for this,” create more specific statements like “there are a large number of purchasers for this product, the existing products in the market are weak, or there are very low switching costs for the buyers,” Lowe says. Then give each a number in a spreadsheet, and start interviewing people.
When you’re interviewing people, don’t ask leading questions. For example, don’t start with phrases like “is it true that…” “You don’t want to state the thing that you want the interviewees to say,” Lowe says. “People are more likely to repeat after you and confirm your bias when you do that.”
Step 2 - Collect and track your nuggets
A nugget is a piece of insight or information you gather from an interview “that validates or invalidates a hypothesis,” Lowe says. Not everything you get from your interviews will be a usable nugget; you want to focus in on the truly important information that people shared with you. “You might have an hour long conversation and get 10 to 15 hypothesis-relevant nuggets,” Lowe says. You want to write each one down and put them in a new page of your spreadsheet, along with the number of the hypothesis they correspond to and whether they validate or invalidate their corresponding hypothesis.
Step 3 - Decide when you have enough data to validate or invalidate
There’s no magical number for each hypothesis that, once reached, tells you that your idea is viable. It’s more about instinct. “The level of significance that you’re willing to accept depends on how important the hypothesis is to your business,” Lowe says. “So if you have a hypothesis about something that’s not a huge lever of success for you, maybe you’re happy with just two or three people validating it. But it’s a game-changing hypothesis, you’re going to want to stress-test it.”
“Over time the list of hypotheses will evolve and you’ll drop some because you know they’re true or false,” Lowe says, “and you’ll be constantly thinking about how to frame a question or an intuition in a statement that can be proven or disproven.”
Read more about taking an objective and data-driven approach to the validation process here.
This post is based on content from a WeWork Labs programming session.
Interested in connecting directly with this mentor? Ask your Labs Manager for help.