Every business idea contains beliefs you have not checked yet: that the problem matters, that a certain person has it, that they would pay to solve it, that your version is better than what they do today. Validation is the work of testing those beliefs one at a time.
The loop below is deliberately simple. You can run one full pass in a weekend, and you will run many passes over the life of a business.
The validation loop
Idea → Assumption → Customer → Evidence → Test → Learn → Improve. Each step exists to stop you from skipping straight to building.
- Idea
- The thing you want to make or do, stated in one sentence.
- Assumption
- A belief inside that idea that could be false.
- Customer
- The specific person whose behavior would prove or disprove it.
- Evidence
- What already exists: what they do, buy, complain about, or avoid.
- Test
- The smallest action that produces a real signal, not an opinion.
- Learn
- What the result actually said — including the parts you did not want.
- Improve
- The change you make to the idea, the customer, or the offer.
Step 1: Turn your idea into assumptions
Write your idea in one sentence, then list what has to be true for it to work. Be blunt. Assumptions written softly cannot be tested.
Example idea: “I will wash and detail cars for neighbours on weekends.”
- People on my street care about clean cars enough to pay someone else.
- They currently either do it themselves or drive to a car wash.
- They would rather pay a teenager on their driveway than drive somewhere.
- I can do one car well in under two hours with supplies I can afford.
- There are at least ten households nearby who fit this.
If one belief being false makes the whole idea pointless, test that one first. Testing easy assumptions first feels productive and teaches you least.
Step 2: Look for evidence that already exists
Before you run a test, collect what is already visible. This is free and it often changes your idea before you spend anything.
- Do people already pay for something similar? Existing competitors are usually good news — they prove the problem is worth money.
- What do people complain about in reviews and community posts about those options?
- What workarounds exist? A group chat, a spreadsheet, or a favour someone keeps asking for all signal unmet demand.
- Has anyone asked you for this before you had an idea? That is the strongest free evidence there is.
If you cannot find a single sign that anyone spends money, time, or effort on this problem today, treat that as a warning rather than an open market.
Step 3: Design a small test with a real signal
A good test produces behavior, is cheap, and can be finished in days. Decide before you start what result would count as a yes and what would count as a no.
Tests that produce real signals
- Ask five people to book a specific time this week and see how many actually commit.
- Do the job once, manually, for one customer at a real price.
- Offer a pre-order or a deposit for something not built yet.
- Post a clear offer in one community and count replies, not likes.
- Run a sign-up page or a simple form and measure who leaves contact details.
Weak signals include likes, “this is cool,” promises to buy “later,” and family enthusiasm. They feel like progress and predict nothing.
Write your success bar as a number before you run it: “If at least three of ten people book a time, I continue.” Deciding afterwards lets you talk yourself into any result.
Step 4: Learn honestly, then improve
After the test, write down what happened before you interpret it. Then separate three possible conclusions, because they lead to very different next moves.
- The problem is wrong
- Nobody cares much. Change the problem, not the marketing.
- The customer is wrong
- The problem is real but this group does not feel it. Try the group where it hurts more.
- The offer is wrong
- The problem and person are right, but your promise, price, or format does not fit. Rewrite the offer and test again.
A test that says no is a success: it cost you days instead of months. The only failed test is one you cannot learn from because you never defined what you were checking.
How do you know when to start building?
You are never certain. You are looking for enough evidence that building becomes the cheapest way to learn the next thing.
- You can describe your customer in one sentence and name real people who fit.
- At least a few of them have shown behavior, not just interest.
- You can state your offer clearly and they understood it without a long explanation.
- The smallest useful version is something you could deliver within a week or two.
When those are true, stop researching and go build the smallest version. Continuing to validate past this point becomes a way of avoiding the risk of shipping.
Coach Kingsley is built to challenge untested claims rather than approve them, so a young founder gets asked “what is your evidence?” before they commit weeks to building.
Common mistakes
Asking “would you buy this?”
Hypothetical yes answers are free to give. Ask for a booking, a deposit, or a first purchase instead.
Testing the easy assumption
Confirming something you already knew feels good but leaves the dangerous belief untested.
Deciding the success bar afterwards
Without a number set in advance, any result can be read as encouraging.
Treating competitors as proof there is no room
Existing paid options usually prove demand. The question is whether you can serve one group noticeably better.
Validating forever
Endless research is a comfortable way to avoid shipping. At some point building is the test.
Run one validation loop this week
One pass, written down. Keep it to a single page so you cannot hide behind detail.
- 1Write your idea in one sentence.
- 2List five assumptions inside it and circle the one that would break everything.
- 3Name the customer whose behavior would prove or disprove that assumption.
- 4Write down evidence that already exists about them.
- 5Design one test you can finish in three days, and write the number that counts as a yes.
- 6Run it, record exactly what happened, and decide whether the problem, customer, or offer needs to change.
Key takeaways
- Validation means finding out you are wrong while it is still cheap.
- Turn your idea into blunt assumptions and test the riskiest one first.
- Free evidence already exists in what people buy, complain about, and work around.
- Real signals are behavior: bookings, deposits, purchases, commitments.
- Set your success bar before the test, not after.
- When your customer, evidence, and offer line up, stop researching and build the smallest version.
Related resources
How to Find Your Target Customer
How to move from “everyone” to one specific person whose problem you understand better than anyone.
What Is an MVP?
Minimum, viable, product — what each word means, with first-version examples for every kind of business.
How to Price Your First Product or Service
Cost, price, profit, margin, and break-even explained — then how to choose a number you can defend.
Browse everything in the Resources hub.
Apply this to your own idea
Learning entrepreneurship works best when it is attached to something you are actually building. Coach Kingsley teaches each step, then asks what you think before it answers.