Skip to content
Kuza Ndoto Academy

Validate Before You Build: Testing a Business Idea in Two Weeks

Most new ventures fail for lack of customers, not lack of effort. A step-by-step plan to find out early.

Personal Development9 min read9 November 2026
A businesswoman presenting data to her team
By Kuza Ndoto Editorial · Photo: Pexels

It is easy to fall in love with a business idea. You picture the logo, the Instagram page, the first big order. Then you spend months and your savings building the product, only to discover that not enough people want to pay for it. This is one of the most common and painful ways new ventures fail, and it is largely avoidable. This article shows you how to test an idea with real customers in about two weeks, before you invest heavily in building anything.

Why good ideas fail without customers

Most failed ventures were not built by lazy people. Founders often work extremely hard. The problem is usually that they worked hard on the wrong thing: a product that solved a problem customers did not have, did not care enough about, or would not pay to fix.

The underlying mistake is treating your idea as a fact rather than a guess. When you say "students will love this", you are making a prediction about other people's behaviour. Until real people act on it, by paying, signing up or changing what they do, it remains a guess.

Validation means collecting evidence, as cheaply and quickly as possible, about whether your guesses are true. It does not prove your business will succeed. It reduces the chance that you spend months building something nobody wants.

Key idea

Your business idea is a set of guesses about customers. Validation is the process of testing those guesses with real behaviour before you spend serious time and money building.

Problem, then customer, then solution

Many first-time founders start with the solution: an app, a product, a service. A stronger approach works in the opposite order.

  1. Problem. What painful, frequent or expensive problem exists? Is it a real problem people already try to solve, or one you imagine they should have?
  2. Customer. Who exactly has this problem? "Everyone" is not an answer. "First-year students living in hostels near campus who cook for themselves" is.
  3. Solution. Only once you understand the problem and the customer should you design a solution. Often, what customers need turns out to be simpler than your original idea.

Working in this order protects you from building a clever solution to a problem nobody has.

Writing falsifiable assumptions

An assumption is something that must be true for your business to work. A falsifiable assumption is written so that evidence could prove it wrong. "People want healthy food" cannot be proven wrong, so it is useless for testing. "At least 1 in 5 hostel students we speak to has spent money on a ready-made lunch in the past week" can be checked, and it can fail.

Write three to five assumptions covering the problem, the customer and willingness to pay. Next to each, write what result would count as a pass and what would count as a fail, before you start testing. Deciding the pass mark in advance stops you from reinterpreting weak results as success later.

Customer discovery interviews

Customer discovery means talking directly to potential customers to understand their problems and current behaviour. It is the cheapest form of validation, and the most often done badly.

The biggest trap is asking leading questions: questions that suggest the answer you want. "Would you buy a healthy lunch delivery service?" invites a polite yes. People are kind, and saying yes to a hypothetical costs them nothing.

Instead, ask about past behaviour, which is far more reliable than opinions about the future. Compare:

Avoid (hypothetical, leading)Ask instead (past behaviour)
Would you use an app for this?Tell me about the last time you had this problem. What did you do?
How much would you pay for this?What have you spent on solving this in the past month?
Do you think this is a good idea?What is the hardest part of dealing with this today?
Would you buy it if it were cheaper?Have you looked for other solutions? What happened?

In each interview, talk less and listen more. Do not pitch your idea until the end, if at all. Take notes on exact words people use, because these often become your best marketing language later. Aim for 10 to 15 interviews with people who closely match your target customer.

Try this

End every interview with two questions: "Is there anything I should have asked but did not?" and "Who else do you know who deals with this problem?" The first often reveals the real issue; the second gives you your next interview.

Smoke tests and the minimum viable product

Interviews tell you about problems. To test whether people will actually act, you need a smoke test: a small, low-cost experiment that measures real customer behaviour before the full product exists. Common smoke tests include:

  • A simple landing page. A one-page website describing the offer, with a clear button to sign up, join a waiting list or order. You measure how many visitors take action.
  • Pre-orders. Asking customers to pay, in full or as a small deposit, before the product is delivered. Money is the strongest signal of real demand. Be transparent about when they will receive it, and refund promptly if you do not proceed.
  • A WhatsApp catalogue or status post. List your product or service on WhatsApp Business or share it on your status and in relevant groups, and count genuine enquiries and orders.

The minimum viable product

A minimum viable product, or MVP, is the simplest version of your product that lets you deliver real value to early customers and learn from how they use it. It is not a poor-quality product. It is a deliberately small one. For a food business, the MVP might be cooking twenty lunches in your own kitchen and delivering them on foot, not opening a restaurant.

Build, measure, learn

Eric Ries, in his work on the Lean Startup approach, describes a loop called build–measure–learn. You build the smallest thing that can test an assumption, measure how real customers respond, and learn whether to adjust your idea. Then you go round again. The goal is to make each loop as fast and cheap as possible, so you can learn many times before your money and motivation run out.

Watch out

Compliments, likes and "I would definitely buy that" are not validation. Count only actions that cost the customer something: time, personal details, or money. Also check local rules before selling food, cosmetics or regulated services, even at test scale.

A two-week validation plan

Here is a day-by-day plan you can adapt. It assumes a few hours a day alongside classes or work.

DayTask
1Write down the problem, target customer and proposed solution in one paragraph each.
2List three to five falsifiable assumptions, each with a pass mark and a fail mark.
3Write an interview guide of eight to ten past-behaviour questions. Line up your first interviewees.
4Conduct three to four interviews. Write notes the same day.
5Conduct three to four more interviews.
6Conduct the final interviews to reach 10 to 15 in total.
7Review your notes. Which assumptions held up? Refine your customer and offer.
8Design your smoke test: a landing page, WhatsApp catalogue or pre-order offer, with a clear price.
9Launch the smoke test and share it with your target customers.
10Keep promoting. Record every visit, enquiry, sign-up and payment.
11Follow up with interested people and try to convert enquiries into pre-orders.
12Deliver a small MVP to your first paying customers, if you have any.
13Ask those customers for feedback: what worked, what did not, would they buy again?
14Compare results with your pass marks and decide: persevere, pivot or stop.

Persevere, pivot or stop

At the end of two weeks, you make a decision based on evidence, not hope.

  • Persevere if most of your key assumptions passed and people paid or committed. Keep going, and plan your next build–measure–learn loop at a slightly larger scale.
  • Pivot if you found a real problem but your solution, customer or price was wrong. A pivot is a deliberate change in one part of your plan while keeping what you learned. Perhaps students did not want lunch delivered, but they badly wanted cheap weekly meal-prep ingredients.
  • Stop if the problem turned out not to matter enough to anyone. Stopping after two weeks is not failure. It is a success: you saved months of effort and money, and you learned how to test the next idea.
Worked example

This is a hypothetical example. The idea and all figures are illustrative.

A second-year student in Nairobi wants to sell healthy packed lunches to students in hostels near her campus. Her key assumption: "At least 1 in 5 hostel students regularly buys lunch near campus and is unhappy with the options."

Interviews (days 3 to 7). She speaks to 12 students. Nine of them bought lunch near campus in the past week, and seven complained about price or long queues, not about health. That is a surprise worth noting: her target customers care more about cost and speed than nutrition.

Smoke test (days 8 to 11). She adjusts her offer to "affordable, pre-ordered lunch ready for collection at 1 pm, no queue" and posts it on a simple WhatsApp catalogue shared in three student groups. Over four days, 140 people view the catalogue and 21 send enquiries, which is 21 ÷ 140 = 15% of viewers. Of those 21, 8 pre-order and pay a small deposit.

Decision (day 14). Her pass mark for the smoke test, set on day 2, was 5 paid pre-orders. She received 8, so that assumption passed. She decides to pivot her positioning from "healthy" to "quick and affordable" and persevere with a second, slightly larger test the following fortnight.

Fall in love with the customer's problem, not with your first solution.

Key takeaways

  • Many ventures fail because they build something customers do not want enough to pay for, not because of lack of effort.
  • Work in order: problem, then customer, then solution.
  • Write falsifiable assumptions with pass and fail marks set before you test.
  • In interviews, ask about past behaviour, not hypothetical opinions, and avoid leading questions.
  • Use smoke tests such as landing pages, pre-orders and WhatsApp catalogues, and build a minimum viable product.
  • Run fast build–measure–learn loops, then decide with evidence whether to persevere, pivot or stop.

Keep reading