Minimum Viable Segment

▶ Watch the original lecture

When you bring an MVP to market, you have built something small, typically a fraction of your total vision. The most important thing is to be able to sell it to somebody repeatedly and be successful when you do. That almost always means finding a minimum viable segment.

Of your grand market (the dream of selling all your stuff to all these people), you find one segment that has consistent needs, based on the four U’s and three D’s, where you think you can take your small product, solve a problem important enough for them, and do it over and over. One of the most important things when starting a company is proving to yourself, and eventually your investors, that you are solving a problem important enough, even if that first segment is relatively tiny. Prove your first idea works (or does not) for a minimum viable segment before you try to blow it out into something much bigger. Then you ignore the other segments for now, and all you do is make that one segment successful.

How big should the minimum viable segment be? It depends on what you are building. Ideally you generate some revenue, but as investors we do not ask about the TAM of a minimum viable segment. The way to think about it: the segment is small enough that if you are right about the problem you are solving, you can dominate it. You have carved out a space where maybe there are not a hundred competitors, you are going right after this thing, solving one particular problem, and you can absolutely dominate it. By dominating it, it becomes viable: you have proven you can succeed with your first idea. And it does not matter if it is your first, second, or third iteration, these are small bets. You are not betting the whole house right out of the gate.

The cluster of demand. The center of the venn diagram is a combination of pain points, budget, product, use case, and channel: how would you sell into that customer profile? The center is your MVS cluster of demand, a term actually coined by Des Traynor, the founder of Intercom, a unicorn out on the West Coast. What you have done is simplified the problem: there is only one channel, you are only talking to one department for their budget, it is a simple product use case. You want it big enough to prove those points, but at this stage it is not about revenue. It is about showing your basic idea works, and that you can repeatably sell it.

Aploi: doing it wrong, then right. Here is an example from our portfolio that first did it the wrong way, then figured it out. Aploi is a healthcare hiring and onboarding platform: nurses, doctors, physician assistants. They had a grand vision with 20 different features, going after nurses, doctors, skilled nursing, home healthcare, senior centers, veterinarians, everything. They spent a lot of time trying to build the product across all of it and did not do well. They struggled to raise money and struggled to get off the ground. Then they peeled the whole thing back and said, “Hiring nurses. Nothing else.” They made that switch before the pandemic (they had tailwinds from it, but the decision came first), and the company started taking off. Now, if you go to their website, guess what is back? Home healthcare, senior living, veterinary care. Because they proved to themselves and their investors that they were onto something, honed the value proposition, then expanded. They are now one of the best-performing companies in our portfolio.

How to find your segment: shoe leather. How do you choose which segment to start with? The best answer is unsatisfying: talk to 200 potential customers in your universe before you do anything, before you spend a dime, before you drop out of school. Get on an airplane, talk to those 200 people, and start doing pattern recognition. It does not cost anything other than time. And when you talk to a customer, you are not just asking “What do you think of my idea?” You are asking “How much would you pay for it?” and “What are the four most critical things in your organization that, if I could solve them for you tomorrow, you’d pay for?” You do not need a complete product. At PayPal we did paper prototypes; you can sit in a Starbucks with paper mockups and describe what you are trying to build. It is guerrilla customer research. With a clickable prototype that has no code behind it, you can refine, test your assumptions, and wash, rinse, repeat until you are confident enough to spend real money on real software development.

Where do you get your 200? That is old-fashioned pounding the pavement. There is no shortcut. You can buy lists, get on LinkedIn, walk into the place where your dog gets taken care of and ask for introductions. So much is online now that it is not hard to generate a list. Conferences are great: buy a guest pass to a conference of vets and start meeting people at the bar. I had a buddy trying to get into the contracting business who started going to architect conferences; he is not an architect, he just bought a pass, met people, and realized there was an underserved segment in construction. Sometimes the ideal approach is not even to dream up a product first: go and have coffee with 200 vets and ask them their pain points. That is even cheaper than inventing a product and testing it.

Two things about validating with customers. First, buying signals can lie. In a company I led called CloudSwitch, right after the 2008 credit crisis (helping people extend their data centers to the cloud), I saw buying signals like I had never seen before, people leaning in, wanting to get started. But nobody was necessarily willing to write a check, and some who did wrote checks for the wrong reason. Grade it yourself: are they saying yes to be polite, or the “Japanese yes” (“Yes, I heard you,” not “Yes, I want to move forward”), or a real yes? Second, watch for bias: if it is only you going to the customer, that is dangerous. Bring a partner to help balance your view. Pick two or three people in the same segment; if two or three say no, that tells you something, but the worst is the people who say yes when they do not mean it.

Padlock the door. Here is one visual to take away. Back in the time machine there was a company called Computer Vision, where John Hirschtick (later founder of SolidWorks) and I worked together. Computer Vision was getting crushed by PTC and spent 50, maybe 100 million trying to build a next-generation product, all over the map with no traction. One of the frustrated project leads in San Diego asked John, “What should we do?” And John said he would padlock the door and put up a big sign: nobody comes in here until they have sat down with at least three customers. So when you ask how you find pain points, the answer is shoe leather: padlock the door, get out of the office, and go talk to a bunch of people.

Build a scorecard. Create a score sheet, and let it evolve. Every time you hear a need, get really specific: ask the customer over and over, “Did I hear the need right? What exactly is this need? If I solved it, what particular problem would I be solving for you?” Then the critical piece: “And what would you pay for that?” Do not forget that question. And if you feel you have them at a place where they say they would pay, ask, “Would you be willing to write me a check in advance?” You will not get this often, but you are onto a winner if the answer is yes. I have started two companies where I had checks from customers before I had any funding.

Segment down to a blatant, critical need. Find a segment where it is not just a nice-to-have but a blatant, critical need, especially in business-to-business. Mobile is so hot that dozens of mobile startups come in saying the same thing: “There are a billion smartphones out there, so if I get 2% of the market with my app at $4.99, I’m a huge business.” The reality is different. Watch how you can zero in. Instead of “anybody with a mobile phone,” say mobile professionals. That is getting into the business world, but there are still a lot of them. Then say field workers, as opposed to desk workers. Then people out in the field servicing goods, as opposed to salespeople. Then servicing medical equipment, as opposed to office equipment like postal machines. Then within hospitals. Then one step further: those machines used for critical care. Now we have gone all the way down to a place where if these machines are not serviced, somebody dies. Now we have a critical need. Notice this path has nothing to do with the size of the business or a particular feature; we found where there is a pain point people genuinely care about. That is what the segmentation exercise should be: getting repeatability and landing in an area where people really care, so you have answered the value question too. It is not just a viable product, it is a valuable product.