How do I validate a startup idea before building anything?
By Bonaventure Ogeto, Founder and CEO, Mctaba Labs
Published
Validate an idea by testing three things in order, before any code: that the problem is real, that people already spend time or money on it, and that they will pay you. Talk to 10 to 15 people who have the problem, then ask for a deposit on M-Pesa. Money or a firm commitment is validation. Compliments are not.
It is easy to validate backwards. You describe the idea to friends, the friends say it sounds great, and that warm feeling gets filed as proof. Then the quote from a developer arrives, the savings go in, and six months later the product is live and nobody is paying.
Your friends were not lying. You asked the wrong question. "Would you use this?" costs nothing to answer, so people answer yes. Validation only starts when saying yes costs the other person something: their time, their reputation or their money.
This guide gives you three tests to run in order, the exact words to use, and a pass mark for each.
What does it mean to validate a startup idea?
It means collecting evidence that people you do not know will pay for the thing, before you spend money building it.
Three kinds of evidence count, from weakest to strongest:
- Time. Someone gives you thirty minutes, then introduces you to two other people with the same problem.
- Reputation. Someone takes you to their boss, their chama or their WhatsApp group and says "listen to this person".
- Money. Someone pays a deposit, signs an order or pays for a manual version of the service.
Everything else is noise. "That is a great idea", "I would definitely use that" and "let me know when it launches" are polite ways to end a conversation.
Which three things do I need to prove, and in what order?
Prove them in this order, because each one is cheaper to test than the next and a failure early saves you the later work.
| Test | The question | How you test it | Pass mark to set before you begin |
|---|---|---|---|
| 1. The problem is real | Do people have this problem often enough to care? | 10 to 15 conversations about the last time it happened | At least 7 describe it without being prompted |
| 2. They already spend on it | What do they do about it today, and what does that cost them? | The same conversations | At least 5 already pay, in money or hours, for a workaround |
| 3. They will pay you | Will they hand over money for your version? | A deposit or a paid manual pilot | At least 3 to 5 pay |
The pass marks are yours to choose. What matters is that you write them down before the first conversation. A founder who decides the pass mark afterwards always passes.
How do I find out whether the problem is real?
Talk to people who have the problem, one at a time, about their life. Do not mention your idea.
Pick a narrow group. "Small businesses" is not a group. "Salon owners in Kisumu with two or more staff" is. If you cannot name ten such people or find them in a WhatsApp group, a market or a Facebook group within a week, that is your first finding: you may not be able to reach your own customers.
Then ask about the past, not the future. People describe what they did accurately. They guess badly about what they would do.
Questions that work
- "Tell me about the last time you had to [do the thing]. What happened?"
- "What was the hardest part?"
- "What did you do about it?"
- "How much did that cost you, in money or in time?"
- "What have you tried that did not work?"
- "Who else is involved when this happens?"
Questions to avoid
- "Would you use an app that does this?"
- "Do you think this is a good idea?"
- "How much would you pay?"
Listen for people who bring the problem up themselves, with some heat. If you have to explain why it is a problem, it is not one for them.
Here is a message you can send to open the conversation:
Hi Wanjiru, Otieno gave me your number. I am trying to understand how salon owners keep track of bookings and no-shows. I am not selling anything. Could I ask you about it for 15 minutes this week, on a call or at the salon?
Keep notes in one sheet: who, what they said in their words, what they do today, what it costs them. After ten conversations the pattern is usually obvious, in one direction or the other.
How do I know people already spend time or money on it?
Look for the workaround. A real problem always has one.
Someone with a real booking problem already has an exercise book at the counter, a staff member who phones clients, or a paid tool they complain about. Someone with a real cash problem already borrows from a mobile lender or a relative. The workaround tells you two things: the problem is painful enough to act on, and what your product must beat.
Put a number on it. If the salon owner loses three bookings a week to no-shows at KES 1,500 each, the problem costs her about KES 18,000 a month. A product at KES 1,000 a month that halves no-shows is an easy decision. If the problem costs her KES 500 a month, no price works.
No workaround at all is a warning. It usually means the problem is not painful, not that nobody has thought of solving it.
How do I test whether people will pay before the product exists?
Sell it before you build it. You have three ways to do that, and none needs a developer.
1. Take a deposit. Go back to the people who had the problem most. Describe what you plan to build in two sentences, give a price and a date, and ask for a deposit to hold a place. Take it on M-Pesa to your own line or a Till. Say clearly that the product is not built and that you will refund everyone if you do not deliver by the date.
2. Run it by hand. Deliver the result yourself with tools you already have. A booking product can begin as you, a spreadsheet and WhatsApp reminders sent by hand each evening. A delivery marketplace can begin as a WhatsApp group and a rider you call. Charge for it. You will learn what customers ask for, which is never quite what you planned to build.
3. Get a written commitment. If you are selling to a business that cannot pay a deposit, ask for a signed letter or an email saying they will pilot it at an agreed price when it is ready. It is weaker than money, and much stronger than a nod.
A script for the deposit conversation:
You told me no-shows cost you about three bookings a week. I am building a tool that sends each client a reminder and takes a small booking fee on M-Pesa, so they turn up. It will be ready on 1 March and will cost KES 1,000 a month. I am taking ten salons for the first group. A deposit of KES 1,000 holds your place and counts as your first month. If I do not deliver by 1 March, I send it back the same day. Would you like a place?
Then stop talking. The silence is the test.
What should I do with the result?
Decide on the date you set, using the pass marks you wrote down.
- Passed all three. Build the smallest version that delivers what those paying customers were promised, and nothing else. They have told you what the first version is.
- Passed one and two, failed three. The problem is real but your offer is wrong. Change the price, the promise or who pays, and run test three again. It takes a week.
- Failed one or two. The problem is not painful enough for this group. Keep the group and look for a problem they raised themselves, or keep the problem and find a group that feels it more.
A failed test is the cheapest outcome in this whole journey. You spent a few weeks and some airtime. The alternative was finding out the same thing after paying for the build.
What does validation change about the product I build?
Almost everything, which is why it comes first.
The conversations tell you which one job the first version must do, so you can leave the other ten features out. The workaround tells you what you are replacing, so you know what "better" has to mean. The deposit conversations tell you the price people accept and how they want to pay, so payments can be designed in from the first day instead of bolted on later. And the manual pilot gives you a list of real requests in customers' own words.
A founder who skips this step tends to ask a developer to build everything, not knowing which part matters. A founder who does it arrives with a short list, a price and ten people waiting. That build is smaller and cheaper, and it has customers on the first day.
Key takeaways
- Validation means evidence that strangers will pay, not agreement that the idea is good.
- Test in order. First the problem, then what people do about it today, then whether they will pay you.
- Ask about what people did last time, never about what they would do.
- A manual version run on WhatsApp, a spreadsheet and M-Pesa is enough to test most ideas.
- Set your pass mark before you begin, so you cannot talk yourself into a yes.
- Two to four weeks and a few thousand shillings of airtime is a realistic budget for all three tests.
Frequently asked questions
- How long should validation take?
- Two to four weeks of evenings and weekends is enough for most ideas. If it is taking months, you are probably polishing instead of asking. Set an end date when you begin and make the decision on that date with whatever evidence you have.
- Do I need a website or a landing page to validate an idea?
- No. A landing page helps when your customers are strangers who find things by searching. For many products sold in Kenya, a WhatsApp message to the right twenty people and a short description they can forward tells you more, faster, than a page nobody visits.
- What if someone steals my idea while I am asking people about it?
- It is a much smaller risk than it feels, because an idea without the months of work behind it is not worth much. The bigger risk is building in secret for a year and finding out nobody wanted it. Talk about the problem openly and keep your detailed plan to yourself if it helps you feel safer.
- Is it honest to take a deposit for a product that does not exist yet?
- Yes, if you say so plainly. Tell people it is not built, what the deposit holds for them, when they will get it, and that you will refund them in full if you do not deliver by that date. Then keep that promise.
- What counts as a failed validation?
- Missing the pass mark you set at the start. If you wrote down that five of fifteen people must pay a deposit and one did, the idea failed in its current form. That is useful. Change the customer, the problem or the price and run the test again.
Sources
- The Mom Test: How to Talk to Customers and Learn If Your Business Is a Good Idea When Everyone Is Lying to You, Rob Fitzpatrick (book, 2013). Checked on 19 November 2026.
- The Lean Startup: How Today's Entrepreneurs Use Continuous Innovation to Create Radically Successful Businesses, Eric Ries (book, 2011). Checked on 19 November 2026.
