Buy Sumizeit infographics

How to Validate a Business Idea

By Sumizeit · Posted on

Learn how to validate a business idea in 30 days. Talk to 15 real people, run one cheap test, and get proof that strangers will pay before you build.

Not sure what to read next? Take the 1-minute personal growth plan quiz Start the quiz →
Share this article

TL;DR

You validate an idea by collecting proof that strangers have the problem and will give up something real to fix it. That means money, time, or an email address. Compliments do not count. Talk to 15 people about their past behavior, run one cheap test with a clear pass mark, and only then decide whether to build, change direction, or walk away.

Most Ideas Die From Indifference

Nick Swinmurn wanted to sell shoes online in 1999. He did not rent a warehouse. He did not buy stock. He walked into shoe stores near his home, took photos of the shoes, and posted them on a simple website. When somebody ordered a pair, he went back to the store, paid full price, and mailed the box himself.

He lost money on every sale, and he did not care. The only thing he wanted to know was whether people would buy shoes they had never tried on. They did. That little experiment became Zappos.

Most founders do the opposite. They get an idea in the shower, feel a rush, and spend six months building it. Then they launch and hear nothing. No angry emails, no complaints, no sales. Just quiet.

That quiet is the most common way a startup ends. CB Insights went through the stories founders wrote after their companies closed, and the reason that came up most in one widely shared report was that nobody needed the product. Running out of cash came second. You can read more about the patterns in why startups fail, but the short version is simple. People rarely fail because they could not build the thing. They fail because they built a thing nobody was waiting for.

So if you want to learn how to validate a business idea, start by accepting one uncomfortable fact. Your excitement proves nothing. Your friends saying "I would totally use that" proves nothing either. Proof comes from behavior.

Start With the Problem, Not the Product

Before you test anything, write your idea as one plain sentence. Use this shape.

A specific kind of person has a specific problem, and today they deal with it by doing something that annoys them.

Here is an example. Freelance designers lose hours every month chasing late invoices, and today they deal with it by sending awkward reminder emails.

Notice what is missing. There is no app in that sentence. There are no features. That is on purpose, because the product is the part you know the least about right now. The person and the problem are the parts you can check this week.

Now pull out the guesses hiding inside your sentence. There are usually three.

  1. These people exist and you can reach them.
  2. The problem bothers them enough that they already spend time or money on it.
  3. They will pay you for a better way.

Eric Ries calls these leap of faith assumptions in The Lean Startup. If any one of them is false, the whole business falls over. Your job is to find the shakiest guess and test that one first. For most ideas it is the second one. Plenty of problems are real and still not painful enough for anyone to open a wallet.

If you like working on paper, Ash Maurya turns this into a one page plan in Running Lean. It takes about twenty minutes to fill in, and it will show you which boxes you filled with facts and which ones you filled with hope.

Talk to 15 People the Right Way

Customer interviews sound boring. They are also the cheapest research you will ever do, and most people do them wrong.

The usual mistake goes like this. You describe your idea with shining eyes, and then you ask, "Would you use this?" The other person is polite. They say yes. You go home happy and learn nothing.

Rob Fitzpatrick wrote a whole book about this trap. In The Mom Test he points out that even your mom will lie to you if you ask whether your idea is good, because she loves you. His fix is to stop talking about your idea and ask about the other person's life instead.

Good questions are about the past, and they are specific.

  1. When was the last time this problem came up?
  2. What did you do about it?
  3. What did that cost you in time or money?
  4. Have you looked for a better way? What did you try?
  5. What happened with the things you tried?

Then stop talking and listen. You are hunting for a few signals. Did they already try to fix it? Did they pay for something? Did they build a clumsy spreadsheet to cope? Those are signs of real pain. If they shrug and say "yeah, it is kind of annoying I guess," you have your answer, and it is not the one you wanted.

Fifteen conversations is a good target. By the tenth or so, you will start hearing the same stories again, and that repeat is the signal that you have learned what this group has to teach you.

Where do you find these people? Go where they already hang out. Reddit threads, Facebook groups, Slack communities, LinkedIn, local meetups. Send a short message that asks for fifteen minutes and makes it clear you are not selling anything. If you get stuck, this guide on finding your first 10 customers when nobody knows you exist covers the same ground, since the people you interview often turn into your first buyers.

One more thing. Do not interview only your friends. They will be kind, and kindness is the enemy here.

Run a Cheap Test That Asks for Something

Interviews tell you the problem is real. They do not tell you whether people will act. For that you need a test where a stranger has to give something up.

Think of it as a ladder. An email address sits on the bottom step. Time sits higher, like booking a call or joining a pilot. Money sits at the top. The higher the step, the stronger the proof.

Here are three tests that have worked for well known companies, and each one cost close to nothing.

The landing page

Joel Gascoigne tested Buffer with a two page website. The first page explained the product. The second page showed prices. If you clicked a plan, you landed on a note saying the product was not ready yet and asking for your email. Enough people clicked the paid plans that he started building. You can copy this in a weekend with any website builder and a small ad budget or a few posts in the right communities.

The demo

Drew Houston could not easily show Dropbox to people, because the hard part was the engineering behind it. So he recorded a short video of the product working and shared it on Hacker News. The waiting list reportedly jumped from about 5,000 people to 75,000 almost overnight. He had his answer before the product was ready for the public.

The presale

This is the strongest test of all. Offer the thing at a discount before it exists, with a clear delivery date and a full refund if you miss it. Ten strangers paying you $30 beats a thousand people saying "cool idea." It is also the test people avoid most, because asking for money feels scary. Ask anyway.

There is also the Zappos approach, where you do the work by hand behind the scenes while the customer sees a finished service. It does not scale, and it does not have to. You are buying information, not building a company yet. Chris Guillebeau fills The $100 Startup with small businesses that began this way, and it is a good reminder that you need less to get going than you think.

Decide What Counts as a Yes Before You Start

Here is where smart people fool themselves. They run a test, get a muddy result, and then explain it away. The ad was shown at a bad time. The headline was off. Those 400 visitors were the wrong visitors.

You can avoid this with one habit. Write down your pass mark before you run the test. Put a number and a date on it.

For example, "If 100 people from my target group visit the page and at least 10 leave their email, I move forward. I will check on the 30th." Or "If I pitch 20 freelancers and 3 prepay, I build the first version."

There is no magic number that fits every business. A $2,000 service needs far fewer yeses than a $5 app. What matters is that you picked the number while you were still calm and honest.

Watch out for vanity numbers too. Likes, page views, and followers feel good and tell you very little. The numbers that matter are the ones tied to someone giving something up. If you want to go deeper on picking the right things to measure, Lean Analytics is the book for that.

What to Do With the Answer

Your test will end in one of three places.

The first is a clear yes. People signed up, paid, or kept asking when it would be ready. Great. Build the smallest version that solves the problem and get it to those same people fast. This piece on how to sell a product people want is a useful next read.

The second is a maybe. The problem turned out to be real, but your answer missed. Or one small group cared a lot while everyone else yawned. This is good news in disguise. Narrow down to the group that cared, change the offer, and test again. Many strong companies came out of a second or third attempt at the same problem.

The third is a no. It stings. But look at what you saved. You spent two or three weeks and maybe $100, and you still have your savings, your energy, and your weekends for the next idea. You did not lose. You dodged the expensive version of the same lesson.

The whole process fits inside 30 days if you keep it tight. Spend the first week on your one sentence and your guesses. Spend the second and third on interviews. Spend the fourth on one test with a pass mark. If you are doing this next to a day job, how to build a successful startup without going broke has more on keeping your costs low while you figure things out.

Build Last

Most advice about starting a business begins with building. I think that is backwards. The founders I admire treat the product as the last step, the thing you earn the right to make after strangers have shown you they want it.

That is really all there is to how to validate a business idea. Name the person and the problem. Ask about what they already do. Run one small test that costs them something. Set your pass mark first, and then believe the result, even when it hurts.

Pick one idea today and write that single sentence. Then find the first person to call. And if you want more ideas from founders who have been through it, start with this list of the best books for founders.

For 15 minute nonfiction book summaries of bestselling books, check out sumizeit.com