Most new initiatives in established businesses follow the same path. Someone has an idea. A business case is written. The business case assumes that customers will want the product, that they will pay a certain price and that the organisation can deliver it. Months later, after the budget is spent, the organisation finds out whether any of that was true.
Founders cannot afford to work that way. They have too little money and too little time. So they turn the process around: before they build anything, they look for the smallest, cheapest test that could prove them wrong. If the idea survives the test, they invest a little more. If it does not, they have lost a week instead of a year.
Established businesses have more money, but the logic is the same. Every euro and every hour spent on an idea that customers do not want is wasted, however carefully the business case was written.
Start with the assumption that could kill the idea
Every new idea rests on a handful of assumptions. Some are safe. Others, if they turn out to be wrong, sink the whole thing. The trick is to find the dangerous ones first.
Write down everything that has to be true for the idea to work. Then ask, for each one: how sure are we, and what happens if we are wrong? The assumptions that are both uncertain and critical are the ones to test. Usually there are only one or two.
For a new subscription service, the critical assumption is rarely whether you can build it. It is whether enough customers will pay for it every month. For a move into a new market, it is often whether customers there have the same problem you solve at home, or whether they solve it in a different way.
Design a test that is small, fast and honest
A good test has three qualities.
- It is small. It costs a fraction of the full investment. Often it can be done by two or three people in a couple of weeks.
- It is fast. You get an answer while the decision is still open. A test that takes six months is not a test, it is a project.
- It is honest. It measures what customers do, not what they say. Asking people whether they would buy something tells you very little. Asking them to sign up, pre-order or pay a deposit tells you a lot.
Some of the most useful tests look almost too simple. A landing page that describes the new service and asks people to sign up. A handful of sales conversations where you present the offer and ask for a commitment. A manual version of the service, delivered by hand to five customers, before anything is automated.
Decide in advance what would change your mind
The most common mistake is to run a test and then interpret the results to fit what you already believed. A test only helps if you decide beforehand what result would make you stop, adjust or continue.
Write it down: if fewer than ten of the fifty companies we contact agree to a pilot, we stop. If more than twenty agree, we build the first version. Agreeing on the threshold before you see the numbers protects you from your own enthusiasm, and from the sunk cost of an idea people have grown attached to.
Make testing normal, not heroic
In many organisations, running a quick test feels like breaking the rules. There is no budget line for it, no template and no committee that approves it. So people either skip the test and go straight to the big investment, or they run the test quietly and hope nobody notices.
Leaders can change this with a few simple moves:
- Set aside a small, flexible budget for tests that does not require a full business case
- Ask for evidence from a test before approving any larger investment in something new
- Celebrate the tests that stopped a bad idea early, not just the ones that confirmed a good one
Over time, the organisation learns that being wrong cheaply is a success, and that being wrong expensively is the real failure.
A simple way to start
Pick one idea that is currently waiting for a decision. Write down its three most critical assumptions. Choose the one you are least sure of, and design a test that could answer it within two weeks for less than a few per cent of the full cost. Decide what result would make you stop.
Then run it. Whatever the answer, you will know more than any business case could tell you, and you will have made your next decision on evidence instead of hope.