Five Skills for Feasibility
I taught the Feasibility Analysis class at USC for a few years. Afterward, a colleague asked me to speak to his graduate class on the topic of feasibility analysis. This gave me the opportunity to do something that I never did over the past semesters – to think about how to express the essentials of feasibility analysis to a new class in one visit with one hour of class time. The following are essential skills as I see them and also the areas where I see many people have trouble. This list is not exhaustive. It’s just the short list of what I recommend for someone starting to judge feasibility of a potential business. I hope that this list and examples help you on your path to evaluating business feasibility.
Five Skills for Feasibility
1. It’s personal.
Business feasibility is subjective. Not just because the process of building a business is a mix of data and art, but because even if a set of people start with the same quantitative and qualitative inputs, they can come to different conclusions. What is judged as feasible for one person may not be for another. We all have different ranges of what we accept as a business that is worth doing. For one person, a “small business” may be fine. For another person, only a highly scalable business is attractive. Both can be good businesses. It’s your decision. Just do not automatically judge scalable businesses as better. These businesses are often also riskier than unscalable businesses, so it’s a balance.
Problems I typically see. “Go big or go home” mentality is encouraged by investors. After all, that’s how they make money. Just remember, there are more “small business” people who take care of all their monetary needs, who employee teams of people, and who produce work that makes their community vibrant. You probably don’t read about them, but they are common around the country. Building a “small” business in the six, seven or eight figures is a good thing. As another professor told me, he once said to a student in his feasibility class that his project was at best a $10M business. The student got depressed. The professor then said that that also meant the student could take home $1M per year. At that point, it’s you decision whether you’d still be depressed.
2. It’s about knowing how to ask questions.
You can spot business opportunities by asking questions. Why is the world the way it is? Why do people do the things they do?
There are a few ways to ask questions as a way of learning about the feasibility of your business. One way is to learn how to interview potential customers and another is to practice being rejected regularly.
A skill to gain in this area is in how to do Customer Discovery interviews with potential customers. Customer Discovery is a process developed by Steve Blank and explained in the book Four Steps to the Epiphany. A core of Customer Discovery is to learn about problems worth solving from potential customers rather than first coming up with solutions that may not be of value to anyone. Part of this process is interviewing, in person, people who may have the problem you are interested to solve.
But how do you interview people? I find again and again that students educated in Customer Discovery and Lean Startup are not comfortable or proficient in the actual interviews themselves. Part of the problem is missing the point of the interviews (to learn about customer problems and what they value). The other problem is in the format of the interviews (leading and hypothetical questions predominate).
To fix these problems, change the way you interview. First, stop proposing solutions and start asking about problems that your interviewees face (related to your area of interest). Do not start with the solution in mind. Do not demo or try to sell the solution you may have thought up. Be open to your interviewees telling you things that are surprising.
Second, stop asking hypothetical questions. This will be harder than it sounds. In normal conversation, hypothetical questions are natural. That is, asking questions that start with phrases like “Would you buy…”, “Should we build…” and other hypotheticals will produce responses that are not helpful to you. The reason is twofold. First, people find it difficult to put themselves in a hypothetical future state and answer your question from that point. What will they really do if you were to build the thing you propose? Will they actually buy it? It’s hard to get good data just by asking, the exception being when talking to an expert and decision-maker in that specific field, like you might find by interviewing enterprise customers. But otherwise, to avoid generating misleading interviews, instead of asking about the future, ask about the past. For example, instead of asking “Will you buy this?” ask “What’s the last time you bought something to solve this problem?” Or “Tell me about the last time you dealt with this. What did you do?” Listen to their stories. Their stories will be different (and more accurate and insightful) than what their hypothetical answers tell you.
The other reason that hypothetical questions are problematic is that people are polite. This is another way of saying that people lie. “Can I really tell her that I’m not interested in what she’s building? She’s clearly worked so hard on it. I’ll just say something encouraging.”
Making these changes will help fix most of the problems that you face when gathering early-stage customer information.
Note that many people think that the way to ask lots of questions is to send out a survey. Before you can send out a survey, Cindy Alverez asks “do you know the questions you need to ask, and do you know the probable universe of answers”? If not, don’t send a survey. In the beginning, don’t send a survey regardless.
If you do these interviews well and have zeroed in on a target early customer group, I expect that you will find convergence in their responses with many fewer interviews than an expected statistically significant sample size. When I got to the point where I could predict what people would say in response to my questions I knew that I had interviewed enough people for now and it was time to move another step forward.
Learn from ideal customers. Who are ideal customers? Steve Blank defines ideal customers as having five qualities. The more of this list (and higher the number) the better.
Ideal customers
- They have a problem
- Are aware of having a problem
- Have been actively searching for a solution
- Have hacked together a solution
- Have or can acquire a budget
Related to this is getting past the fear of talking to strangers, which you will have to do repeatedly. For this, on the first day of feasibility class I have students start what I call the “Inoculation Assignment.” This is 30 days of going out and getting rejected for a request from another person. Variations of this are sometimes called “rejection therapy.” Do it and cure your fear of talking to people.
Problems I typically see. Believing that you already have all the answers up front. Not wanting to spend the time talking to customers, claiming that other companies don’t do this (insert common ones such as Apple, Facebook or Twitter here), and that it’s all about vision. Understanding that the above style of interviewing may not work in every culture. For places where it is unheard of to talk about problems, you will need to gather insights in other ways, including through stories and testing actions.
3. Learn to build an MVP.
OK, as a first step, learn what an MVP is. MVP stands for “Minimum Viable Product” and is a term popularized by Eric Ries, author of The Lean Startup. The point of building an MVP is to maximize learning per unit of effort. You will determine the type of MVP to do based on what you are working on, what your skill set is (what you can actually build) and how high your threshold is for testing your ideas.
Of course, you do need to figure out what you are trying to test and what counts as a “success.” Otherwise, you’ll keep moving ahead no matter what. Here are seven examples of ways you could build an MVP:
And here are the pros and cons of each:
Most of the people I meet who talk about MVPs do not know what they are. Partly it’s because it’s hard to wrap your head around the MVP’s purpose. The official definition is a little dense. Never has one sentence so confused people.
Here’s my new definition.
When it comes to MVPs, I see people have problems with knowing what they are trying to learn. That is, if you start without preset guidelines for what you are trying to test and what qualifies as a go / no-go decision, you can easily fool yourself. Everything will look like success. Five people clicked on the smoke test? Success! I guess… A conversion rate of 10%? Success! I guess… Now, those numbers could have truly been marks of success if you thought in advance about why. But if you never think about what determines success at this stage for you, you will likely fool yourself into moving ahead no matter what.
Later on, it will be more helpful for you to gather on actual usage.
What metrics matter for your business? Depending on your business type, different metrics will matter for you. A sample of metrics that you might track (some of these are related): customer acquisition cost, lifetime value, retention, aspects of growth such as referral rate and upgrade rate…
Problems I typically see. No thought about this or go/no-go qualifier in advance of “testing.” Not understanding what drives the business and therefore what should be tracked.
4. Understand basic accounting, including cash flow.
Understand what startup costs are for your potential business. If you need significant capital just to get started (not necessarily a bad thing) do have access to that capital at your current stage? Is it worth trying to raise it or use your own savings for this business? What is the likely payback period? What needs to happen for you to have an exit or build a sustainable business?
Understand what timing of payments you will experience. For example, if you do the work to sell to a customer, how long does it take you to get paid? For some businesses, like a cash-only concierge health care service, the service provided and the payment are close together. For other businesses, for example, an apparel business, the production of inventory happens months before the clothes reach and are paid for by customers. In tech, it is this timing of payments that often impacts hardware companies, which need to buy or produce components, assemble and the ship before being paid. Of course, there are ways around this, such as pre-orders or crowdfunding.
Problems I typically see. Thinking that accounting is boring and therefore something to avoid. Not seeing how basic accounting knowledge can help you understand what’s happening (or going to happen) in your business.
5. Think hard about the channels through which you reach your customers.
I see many early-stage companies that have most of a team together, who can build, who have unit economics figured out, but who do not know how to reach their customer base. Lots of companies die not from producing bad products, but from not understanding how to reach their customers.
Problems I typically see. Being fooled into thinking that good products don’t need to market themselves.
Bonus. Know that things change. Sometimes you can see the direction in which things will change.
As I told my feasibility classes, back in the 1930s when the change in recording technology (microphones that were starting to finally record the human voice accurately) preceded a large recording industry, you could predict that years later, something big is going to happen in the music industry. You won’t know which specific company will win, but you could know that there will be growth.
If back in the early 1980s you looked at likely trends in component size and battery life, you might have also predicted that mobile telephony would be popular. Instead, in the early 1980s, when AT&T hired McKinsey to advise them on the future of the cell phone industry, the consultancy advised exiting the market. While AT&T had innovated the cell phone handsets and created wireless networks, the devices were still large, expensive and with short battery life. McKinsey estimated that the market for handsets was only 900,000 units. AT&T exited the market but realized its mistake. In the 1990s AT&T bought back the cellular networks it had sold off (paying $12.6 Billion).
Problems I typically see: Using the way the world is now to project how the world will be. Using market size for a comparable to project market size for a new service. In other words, using the size of the taxi market to predict how big companies like Uber or Lyft could be.
This is not an exhaustive list. You can’t do everything in a feasibility analysis. I don’t focus much on competition or market size at the early stage, for example. There will be enough gaps in knowledge or predictability there to lead you astray before you have real data. I’d rather that my students just get moving on learning.