Customer Driven Guide

Chapter 2 - Customer Discovery

Mike Wilner works in Business Development on the Early Stage Startups team at Amazon Web Services. He co-authored the book Oversubscribed: A Founder’s Guide to Seed Raising. It’s been read by thousands of startup founders and used by Accelerators and Incubators to educate founders on seed raising. Mike has founded several companies and also served as an early stage startup consultant. He received a B.S. in Math and Business from Washington and Lee University. 


In early stage startups, you often don’t have enough data to answer hard questions. At least, not with any kind of statistical significance. This makes product development extremely difficult. How do you validate direction? What do you test? And what does progress look like anyhow? 


In his current role, Mike spends a lot of time talking with founders who are wrestling with these questions. In this interview, we dive into de-risking decisions, growth vs. progress, staying close to the customer, and successfully navigating small loops to faster, sustainable product building. 

Stuart Balcombe: Small loops apply to everything, not just products or early validation. You can apply them at any stage of a business, right? 

Mike Wilner: Yeah, I see this a lot at Amazon. The startup team launches a lot of different programs, projects, and experiments. So much of it is, “Can you do something that works? Is it repeatable?”


Then you start repeating it, and next thing you know, you're building a team around it. Finding anything that works really well repeatedly, gives customers a good experience, and that you can get value from, that’s product-market fit at a small scale. 


Finding anything that works really well repeatedly...that’s product-market fit at a small scale. 


To find this, you have to be ruthless about killing things and understanding when they don’t work. I find a lot of founders undervalue that. 

What is your barometer for “working really well”? 

I was having a conversation with a founder about this the other day. I said, “You need to find the customer who is using your product and you are a part of their lives. If you were to go out of business, they would be extremely disappointed. It wouldn’t just be really painful for them to go back to the status quo—it would be a pain in the ass.” 


“Working really well” is when your customer would be upset if you went out of business.


For example, I was talking with a firm who had buggy software. Logging in and accessing certain things wasn’t working well, and it was preventing their users from getting to customer success.


“Working really well” is when your customer would be upset if you went out of business.


They had different customers in different segments, and I said, “Hold on. Who's the most pissed at you right now?” And they had one customer that’d been hounding them about the issues. I said, “That's your customer. He clearly sees the value in this.”

Another great example is Slack, right? Whenever Slack is down, there are memes about it and everyone takes to Twitter saying, “Well, what are we going to do the rest of the day?” It’s so integral to their customers’ day-to-day.  


Right.

Do you have an example of someone who’s gone through this small loops process—picking something small that’d be a success, then testing and validating it?

Yeah, one founder I was talking to recently has a vision for taking food and using it as medicine. She wants to build a meal kits business for specific ailments. 


In the early days, her team had all these ideas for different types of boxes they could create. They were doing the math around how many SKUs they would need for the different combinations, and I told them, ”I'm not concerned about your ability to create new boxes. That's not a big risk.” 


At this point, the biggest risk is: can you find your ideal customers? Do they want this? And are you able to deliver it to one of these segments? 


At this point, the biggest risk is: can you find your ideal customers?


The first thing she needed to do was nail one box—one subscription. She needed to show she could create the product, do it all on a small scale, and then get people to really love it. And she was able to validate that. They understand the channels. They know how to go get customers, serve them, delight them, have them feel really good about it, want it more, and want to refer it to their friends.


They figured this out with one box, and now they have a playbook they can run with for different ailments. 


And so what's interesting to me is they raised 2.5 million dollars with a little bit of traction around this one box, and it’s not even impressive traction. 


But they can do this because it's not about revenue. I actually have no concerns about their ability to get to a million in annual revenue. They know where they're going to get their customers. They know where they're going to go to market. They've shown that they can repeat this. 


It's taken six to twelve months of going through this to figure it all out, but they know how they're going to be able to grow the business once they decide to. They've validated that and didn’t even have to be totally in the market to do it. That's just one example.

You mentioned traction. What makes it so important to de-risk the steps you take to get traction? You see a lot of founders take huge risks like fundraising, ad spending, growing a team...all so they can grow. But I guess growth and traction are not necessarily the same thing?

There is a difference between progress and growth. Progress is measured by how much you de-risk. This applies differently to software businesses vs. deep tech and hardware businesses. 


Progress is measured by how much you de-risk. 


With deep tech and hardware, “Can you build it?” is the biggest risk. And so the place to start with those businesses—assuming you know there is an application for what you want to build—is to try to build it.


With B2B SaaS, “Can it be built?” is not a risk. The answer to that is yes. There is no B2B, SaaS, or any kind of regular old software where, “Is it possible to build technically?” is the actual risk. You simply need good developers and good product designers to build a product. 


With B2B SaaS, “Can it be built?” is not a risk. The answer to that is yes.


The biggest risks for B2B SaaS are: Do people want it? Do you know what to build? Will they be able to adopt it? 


It’s about always identifying the biggest risk and uncertainty, and tackling that aggressively. The faster you can de-risk things, the faster you're making progress. Founders often underestimate how quickly they can make substantial progress by thinking smaller and de-risking. 


The faster you de-risk things, the faster you’re making progress. 


Getting one pilot customer, depending on what you're doing, can be a massive amount of de-risking.

Particularly with software, this is where a lot of founders get themselves into trouble. Especially if they’re non-technical or new to software, they assume building software is the hardest thing. But if you try to validate or de-risk an assumption by building software, it's going to take a long time. However, you can de-risk, “Can you get your first customer?” without building the product. 

What are the other big misconceptions you hear from founders when you talk with them about de-risking? 

Founders struggle with de-risking when they roughly state unknowns and go after them, instead of stating a specific hypotheses. Because really, when you think about risks, risks come back to hypotheses. You have a hypothesis and you don’t know if it’s correct. 


We can build this product; we can build this medical device; people, ops managers will be able to onboard their employees onto this; they'll be able to use it...these are all hypotheses. 


And then you have to prove or disprove them. One of the big misconceptions—and this is where you get into two way doors—is that proving a hypothesis wrong is bad. Proving a hypothesis wrong is actually really great progress. 


Proving a hypothesis wrong is actually really great progress. 


Because what's the alternative? Sweeping it under the rug and finding out later an assumption you glazed over is wrong? That fucks everything else up. Figuring out what you're wrong about is really important and having the intellectual honesty to be able to say, “I was wrong about this,” is important. 


Often, if you don't explicitly state your hypothesis, it's hard to know if you're right or wrong. You end up in this cycle of customer development or feedback where you kind of forget what you're trying to answer. You're just dealing with noise. 


The speed at which you prove your hypothesis wrong is very directly tied to the speed at which you’re making progress. 


Whereas if you're really honest, and you clearly state what your hypothesis is, it's possible you could have one conversation and disprove it. You’re like, “Shit, I was wrong. Let's figure out what the new hypothesis is here.” The speed at which you prove your hypothesis wrong is very directly tied to the speed at which you're making progress.

It's interesting that most founders, when they're talking about de-risking and product validation, talk about trying to validate their idea. That opens you up to a lot of bias. Whereas if you flip the question and say, “How can I disprove this?” it typically takes less data points. 

Or if you’re testing a hypothesis in one vertical, and disprove a really important hypothesis there, maybe that just isn't a good vertical. Maybe you need to test it in a different vertical. 

Yeah, it’s important to have multiple hypotheses and separate the riskiest one. This is the litmus test I use: the riskiest hypothesis is the one where if you’re wrong, nothing else really matters. 


The riskiest hypothesis is the one where if you’re wrong, nothing else matters. 


If you're right, it's a big fucking deal and let's celebrate it. But you should get those answers quickly. 


One thing that comes to mind, and this is something I really enjoy learning about at Amazon, is the principle “Be right. A lot.” One of the biggest lessons I learned is that you don’t have to be right all the time. You just have to be right more than you're wrong. And when you're wrong, you have to realize it quickly and acknowledge it. 


...you don’t have to be right all the time. You just have to be right more than you’re wrong.


The people who are right a lot? They just reduce the amount of time they're wrong because they aggressively work to disprove their assumptions, and they find out where they’re wrong quickly. They're honest, and they're not scared to figure that out. 


At Amazon, in conversations with my boss, I’ll have pretty high conviction about something. And I'm getting excited, getting some momentum, and then my boss will say something data driven and provide a reason why I'm wrong. And now I’m so quick to go, “Yep bad idea.” And that's okay. 


You have to be open in the presence of information that disproves your idea. You want to be right in the long run. But being right in the long run means being wrong a lot along the way. Feeling like that’s not a critique on yourself is part of learning to be right in the long run.

Yeah, I think that's a great lesson in itself, right? That can really change how founders think about decision-making their own business. 

Right, it’s about intellectual honesty and just being willing to be wrong often. That's really important.

You mentioned that in order to be right a lot, you also have to be wrong a lot, learn quickly, and adjust when you’re wrong. This is something that's directly tied to customer discovery and being customer focused. How much time should founders spend trying to de-risk assumptions with customers?

Most of the time. That's why CEOs do all the sales early on. The closer you can be to the customers, the better. If you don't create mechanisms that help you stay close to the customer, then your product is going to degrade over time.

Going the other way, what would be a red flag? For example, if a CEO hadn't talked to a customer in X time, you'd start to be worried about them.

I don't think it's week over week. It can come in sprints. 


For example, if you're a CEO, and you've got to fundraise, you may spend two months where you're only focused on fundraising. But before you go and fundraise, you should spend enough time close to customers where you de-risk a lot of those customer-centric assumptions. 


The closer you can be to the customers, the better. 


Regardless, you need to have some mechanisms in place so your team is staying close to customers. One thing that’s really dangerous is, good CEOs are good at staying close to customers in their early days. But then things like fundraising and hiring come up, and they move further away from the customer. You have to work to create mechanisms so being customer-centric is core to how your company operates. 


If you don't create mechanisms that help you stay close to the customer, then your product is going to degrade over time.


And that’s something pretty cool at Amazon. Customer-obsession is a leadership principle and culturally everything is about being customer obsessed. There are a lot of mechanisms in place that force people to remain close to customers and think of customers in every product decision. You have to work really hard, especially as your organization grows, to build those mechanisms in. So even if you're not doing customer interviews, you are close to the customer in some way.

What do you tell founders who feel too busy to do customer research? Or who want to spend time building things on intuition. What would you tell that person? 


To build those mechanisms. 


And you know, intuition is a funny word. Ultimately, I don't think intuition is a thing. The founders who are lauded as having great intuition, they don't just sit in a room and have a good intuition. 


Successful intuition is just customer obsession. 


They have good intuition because they spend all their time talking to customers. Successful intuition is just customer obsession. 

They’re filtering things the customer say into the actions they take. 


Yeah, and I will say intuition without customer obsession is a first time ignorant founder who is not going to succeed. 


No investor worth their salt believes in the idea of a true visionary who is not customer-centric and customer-obsessed. That is not a thing that exists in the world, period. 


No investor worth their salt believes in the idea of a true visionary who is not customer-centric and customer-obsessed. 


Steve Jobs is the classic example. He was absolutely customer obsessed. He talked to people all the time about how they experience and want to interact with products. Then he had certain visions for how they could do that. He didn't just build what people asked for. 


Customer-obsessed does not mean you just built what people ask for. It means you are deeply engaged in what people need. 


And so I'd guess Steve Jobs was one the most customer-obsessed people ever. And that's why he was able to build the things he built.

You mentioned Amazon Leadership Principles and two-way doors have shaped your thinking around high-risk decisions. What else? 

Another thing I talk about with aspiring founders is keeping your opportunity costs really high. Meaning you should really love your day job. And you should also really want to start a company. But if you really love your day job, you’re going to keep the bar very high for what you have enough conviction to actually dive into full-time. 


And then I’ve also been thinking about how you kill ideas. 


I've been looking at how different startup studios make what I call go/no go decisions where they're deciding whether they’ll double down on something or scrap it, so the very valuable talent they have can work on something else. Because time is the most precious asset that they have. 



This interview has been edited and condensed for clarity.