Jim Gray

Data science, design for usability & outcome optimization.
Working on marketing tools to multiply capabilities of small teams.

Intro to Effective Customer Development

Customer development is, broadly speaking, the process of doing directed work to improve your understanding of your customer. We want to know about what problems they have, how they see and interact with those problems, and what their values are.

You can make some good guesses based on experience, but people will always surprise you. There's often a huge disconnect between how businesses conceptualize their products and their sales process, and what is actually happening from the customer's perspective.

By better understanding the customer's perspective, we hope to make better products, do more effective marketing, and in general begin to make more money while doing more rewarding work.

Using customer development to guide your product development should be the polar opposite of the "let's build this, then see who wants it" approach. Products don't sell themselves, even if your solution really is superior to existing options.

There are benefits to this process at all points in the product cycle — not just pre-launch. It can help you explore issues like what is and isn't working in your marketing, what features are really worth developing, or how to improve your pricing and product positioning.

Why customer development?

To begin with, most market research is crap. Most of it really is little better than the quickie pie chart infographics that get shared around on Facebook. Studies are structured according to what won't take up a lot of time and effort, or to use the expertise people already have instead of the methods that are really appropriate to the questions being asked.

The problem is that experimental rigor is what keeps the results tied to reality. When the result is judged based on how impressive the PowerPoint looks, or on whether it's supporting what people already wanted to hear — the result is often little better than flipping a coin.

Perversely, people tend not to act on their market research anyway. People know that they should be doing research, so they go through the motions, but they don't know how to do it effectively. And they also rarely know how to use the results to guide their actions towards better outcomes.

And companies often do their customer research after development, in the time leading up to launch. But at that point, sunk costs, limited resources, and internal politics mean that you probably can't make drastic changes, no matter how badly they're needed.

End result, you're running around blindfolded in a jagged landscape of customer who really don't care about your company — with numerous competitors waiting to walk over you when you end up crashing into reality.

Good research won't solve everything, but it means you'll at least have valid data to work from, and a broader view of the arena you're fighting in. And if much of your competition is still running blind — well, I'm sure you get the picture.

Two assumptions:

  • I'm mainly covering customer interviews here. They're not the only tool, but they often maximize your information gained for the time involved. They're also one of the most under-utilized tools.
  • You're at an early stage vs. trying to improve an existing product. If you already have paying customers, you should usually do some work analyzing your customer data first instead of starting from a blank slate.

Tools like focus groups and customer surveys are frequently misused, but they can give you actionable information. It just takes a lot of finesse and expertise, and there are a lot of ways to screw up the results without realizing it.

However, it's not that hard to learn to be a decent interviewer, and customer interviews are one of the most efficient ways to keep in touch with what's really going on from your customer's perspective.

Okay, so customer development is good…

It's easy enough to do effective customer development once you get in the habit, even if you're new to the research game. You just have to hold yourself accountable, and avoid some common pitfalls, and accept the results you find whether or not they match what you hoped to find.

You can always investigate further if you suspect that the initial results are misleading. The fundamental idea of research is that you'll use what you know to gather information that will help improve what you know. So do that.

You should start doing customer research before you start building anything, unless it starts out as a personal project. But even then, you'll benefit greatly if you start doing customer research as soon as you start trying to make the transition from personal project to commercial product.

Due to the nature of commercial projects, large changes are often impossible to push through after the very early stages of the process. So we need to make an impact as far upstream as possible. Early research also helps us create better products, more efficiently, with greatly increased odds of market success.

We're also going try to use reasonably scientific methods to ensure that our data is both useful and valid. This helps us to work methodically towards a useful understanding of what our customers need and will respond to. Bad methodology means getting results which are heavily biased, if not completely unrelated to the truths we're trying to uncover.

And yes, you can do this even if you're not a social butterfly. You may find the first few attempts utterly nerve-wracking, but you can do it. Really, approaching people is the hard part. Once you get into the conversations, you're talking about an issue that both of you care about. And if you do it right, they'll do most of the talking.

Your customers matter. Everyone else, not so much.

Talking to random people to learn about your customer, instead of talking to your customer to learn about your customer, is so off target that it's not even wrong. This will end up either confirming your biases or giving arbitrary results, neither of which is a positive outcome for your business.

Approaches like posting public surveys on Facebook, or hiring students to stand around on a street corner with a clipboard, or even trying to copy what your competition is doing — get that stuff out of your head right now, because it's awful and you're never doing it again.

To learn about your customers, you have to find and talk to your customers. And we're not talking broad demographic groups. You need to get specific. I'm talking laser-focused. We're shooting for the 10% that will be your most rabid fans, not the other 90% who will just be satisfied customers.

If you're doing it right, most of this should match up.

  • What problem hurts really bad
  • Why it's a problem, what goal it's interfering with
  • How they experience the problem
  • How they think about the problem
  • How they talk about the problem
  • How they currently deal with the problem
  • Where you can find the target customer
  • What approaches they are receptive to

If what you're hearing isn't very closely aligned after number of interviews, then you should use what you learn from the first round to start over with better targeting. In fact, assume that you'll need to do this at least once.

Also: friends and family don't count. Ever.

They want you to succeed, so they're going to humor you when you need to dig up hard truths. Or on the other side of the coin, they might doubt you when the customer data strongly supports bringing a product to market.

Besides which, you're in this to learn how to find and sell to and satisfy your prospective customers. Learning how to effectively hunt down and engage your target customer is customer development.

Don't ask the wrong questions.

Even if you do all of the above right, you need to be asking the right questions. Some types of questions are just ineffective at digging up information. Other types, they'll get responses — but people will give you the responses they think you want, or they'll answer based on an idealized version of themselves.

Just to give you a few: their income, their weight, how often they exercise, how healthy their diet is, flossing, and most anything else that features prominently in New Year's resolutions. Avoid asking questions like those, at least not without further probing to get to the real meat of the issue.

Biased questions are a huge trap. Anything that assumes a solution, or that imposes a viewpoint, is completely at odds with getting the customer to tell you about their viewpoint. People will naturally tell you what they think you want to hear if you do anything to express an expectation. Or if there's a cultural expectation that says what the "right" answer is.

You also need to avoid questions that are going to be a conversational dead end. These questions often suggest a "yes or no" or "multiple choice" response. They just don't go anywhere, and they often only get you information that is rooted in your preconceptions. Whenever you use a question like this, it should be leading into a stronger question.

Ditto for soft-serve questions. If you go asking a politician if they want to create jobs and be tough on crime — well what answer do you really expect? Questions like this won't tell you a thing, they're just a waste of time.

Don't even think about asking whether they'd buy your product, or how much they'd pay for it. You're doing customer research. You are not selling your product. You are not even supporting a particular solution yet. Asking these questions is one of the easiest ways to torpedo your research.

Always remember that your role during customer development is to be the student, not the expert, and definitely not a salesman. Get them to teach you about their problems, their views, their values.

The customer is the expert. You're the one trying to learn about their life and their problems. Behave accordingly, both with how you approach them and with the sort of questions you ask. Worry about selling yourself and your product later, after you've done everything you can to learn about what they need and value.

So what sort of questions should you ask?

What you want is open-ended questions that will draw people out and get them to tell you rich, detailed stories about their experiences and about how they see the problem.

You also want to draw them into giving you concrete examples — when did you last exercise, what routine did you use, not "do you exercise often."

When in doubt, dig deeper. Ask why. Ask how. Ask them to tell you about the last time it happened. Or, say nothing and wait for them to fill the silence with more of that information you need.

If you're talking more than 25% of the time, you're doing it wrong. You're here to listen. Say just enough to guide the discussion and to get them to keep telling you about themselves, what problems they have, and how they currently deal with them.

If you talk less, you'll have less opportunity to create bias. They'll also have more opportunity to talk. Both of these things are wildly positive for you. Talk as little as possible while keeping the conversation moving and on-topic.

  • Ask about specific instances, not generalities
  • What they do is far more important than what they say
  • Always dig deeper to find the underlying reasons

When in doubt, be a mirror. Use their language to ask for more information about what they just said. It's an exceptionally effective way to draw people out while minimizing the risk of bias.

Customer I kind of resent my manager. You Why do you say that you resent your manager? Customer Well, he's always waiting until the last minute to ask me to come in on the weekend. And then there's those TPS reports — I just can't stand dealing with TPS reports. You Can you tell me about the last time you did TPS reports?

Also, you're not here to judge. Empathize, yes. But get used to the fact that people are going to tell you some fairly strange stories about how they currently get things done. The business world is held together with Excel and bubblegum. You are the student, they are the expert.

As to pricing concerns — while you shouldn't ask directly, you can certainly get them to tell you how about their current solution, and about the amount of time and resources that it's taking up. You can work out what that means in terms of monetary value and get a much better picture of what you could charge than by asking them to make up a number.

Interview tricks and tips

Do your interviews face to face. Failing that, do them over Skype or on the phone. But face to face is best. You just can't have much of a conversation over email, and it's much easier to draw people out if you're talking in person.

The less you talk, the less opportunity there is for you to create bias. Sometimes the best question is to say absolutely nothing. People will often volunteer more information in order to fill the silence.

Whenever possible, use their words. It shows them that you're listening to them and it makes it much easier to engage them in conversation. Their words for describing the problem are also something you absolutely want to capture, it will be a huge resource when it comes time to do marketing.

Ask questions like you're a complete novice. Don't make assumptions about how people do things. You want things like exactly how they do simple tasks, even if you think you already know how they'd do it. This is some of the most valuable information for problem discovery.

Don't do checklist interviews. You're looking for open-ended exploration of a topic. If you come in with a list of ten questions and try to knock them off one by one like a high school news reporter, all you'll get is a reflection of your initial assumptions. Have a conversation.

Avoid distractions. You want a quiet environment where they'll feel comfortable. This is almost never your office. Your office establishes you as the one in charge, but you want them to feel like the expert. Go to them, or else meet them somewhere neutral like a coffee shop. But make sure it's a relatively quiet and low-traffic setting.

Use a voice recorder. You don't want to waste attention on notes, and you don't want to be constantly looking away from the interviewee either. Also, it will do a much better job of capturing the session than your notes could.

You can use your phone as a recorder if you want. There are arguments in favor of this being less threatening. I prefer to use a professional recorder that captures higher quality audio — if you're doing it right, they'll be looking at you anyway and not at what's on the table.

By way of specific recommendations: the Zoom ZH1 is ample for $99, and you can have your recordings transcribed by CastingWords for $1/minute if you feel the need. Make sure to label and back up your recordings!

As to the number of interview subjects, you want to get at least five to ten people who are telling you very similar stories. If it takes more than that to converge because you're still trying to figure out how to tell who is and isn't your target customer, then it takes more than that. Iterate as necessary.

And again — no selling, no promoting your solution, and no assuming the solution. You are the novice, they are the expert. Act the part and let them give you the information you need to make a successful product.

Happy interviewing!