One of the most talented people on my team once asked me how I had become so decisive.
The honest answer is experience. Many decisions that now appear intuitive to me are the result of situations I have encountered before, mistakes I do not want to repeat and patterns I have learned to recognise. After you have seen the consequences of delaying a difficult decision, you become more willing to make it earlier. After you have watched an avoidable problem become expensive, you learn to act while the problem is still small.
However, decisiveness does not mean knowing with certainty that your idea is correct. It also does not mean attempting every idea at full scale. No company has enough time, money or people to pursue every possibility with the same intensity.
The founder’s task is to find a way to test many ideas without betting the company on each one.
Strong opinions should produce cheap experiments.
Decisiveness is not the absence of doubt
People sometimes interpret decisiveness as complete confidence. They imagine that a decisive founder sees the answer clearly, makes the choice and moves forward without uncertainty. In reality, many important decisions are made with incomplete information.
The founder may believe that a new customer segment will buy, that a different price will improve conversion or that a particular partnership will create distribution. The belief may be strong, but until the market responds, it remains a hypothesis.
A strong opinion is useful because it gives the company a direction. Without conviction, the team can spend months discussing every possible option without learning anything. Yet conviction becomes dangerous when it requires the company to make an expensive commitment before gathering evidence.
The solution is not to weaken the opinion. It is to lower the cost of being wrong.
Instead of debating indefinitely whether an idea will work, ask what the smallest credible version of that idea would look like. What can we do within a week, with a limited budget and a defined group of customers, that would give us useful evidence?
Decisiveness means deciding what to test, how much you are willing to risk and what the result will cause you to do next.
You cannot try everything at full scale
Founders encounter more ideas than their companies can execute. Employees have suggestions, customers request features, competitors introduce new products and investors recommend markets that appear attractive. The founder also has personal ideas, many of which feel urgent because they arrive with the excitement of discovery.
If every idea becomes a major project, the company will become a collection of unfinished experiments. Engineering attention will be divided, marketing will communicate several conflicting stories and customers will struggle to understand what the company does exceptionally well.
You cannot try every approach at full scale, but you can often find a smaller way to test the underlying assumption.
If you believe a new service will generate demand, you may not need to build the complete technology immediately. You can offer the service manually to a small number of customers and observe whether they pay, return and recommend it.
If you believe a new market is attractive, you may not need to establish a full office and hire a local team. You can speak with customers, find a distribution partner, close the first few contracts and learn which regulatory or operational assumptions were wrong.
If you believe a marketing message will improve acquisition, you do not need to redesign the entire brand. You can test the message with a defined audience, measure the response and compare it with the existing approach.
The objective is to preserve the ambition of the idea while reducing the cost of discovering whether it deserves more investment.
An experiment must be designed to teach
Many activities are described as experiments even though they are incapable of producing a clear lesson.
A team launches a campaign without defining the expected result. A product feature is released to every customer without identifying which behaviour should change. A new employee is hired to “drive growth” without agreeing on what meaningful progress should look like. After several months, the company has spent money but cannot say whether the original assumption was correct.
That is not an experiment. It is activity accompanied by hope.
A useful experiment should answer five questions before it begins:
- What do we believe?
State the hypothesis clearly. For example: “Businesses in this segment will pay for this service because the existing process exposes them to a significant financial risk.” - What is the smallest credible test?
The test must be inexpensive, but it must still resemble reality closely enough to produce useful information. Asking people whether they like an idea is cheaper than asking them to pay, but payment provides stronger evidence. - What will we measure?
Decide whether the important result is conversion, revenue, retention, completion time, customer satisfaction, cost reduction or another observable behaviour. - How much time and capital will we allow?
Every experiment needs a boundary. Without one, an unsuccessful test can quietly become a permanent project because nobody wants to admit that it failed. - What decision follows each outcome?
Define what will cause the company to invest more, modify the idea or stop. If every possible result leads to continuing the project, the experiment was designed to protect the opinion rather than test it.
Cheap does not mean careless. A poorly designed test may save money initially while producing misleading information that causes a much more expensive decision later.
Manual work can purchase knowledge
One of the most useful ways to run a cheap experiment is to perform the process manually before automating it.
Manual execution allows the founder to experience the customer’s problem directly. You see where information is missing, which steps create anxiety, what exceptions occur and which part of the service customers value most. This knowledge can prevent the company from spending months building software around an inaccurate understanding of the workflow.
Some of the most valuable early work does not scale, because its purpose is not yet scale. Its purpose is learning.
However, manual work needs a time limit and a scale limit. Once the company understands the process and sees repeated demand, continuing manually can become an operational weakness. What began as an experiment must eventually become a system.
The sequence matters:
First, perform the work closely enough to understand it. Then simplify the process. After that, automate what is stable and repeatedly valuable.
Technology should scale learning that has already occurred, not conceal a process the company does not yet understand.
Increase investment as evidence increases
At Eazipay, several of the things we now pursue with conviction did not begin as enormous company-wide bets. We tried approaches in smaller forms, studied the outcomes and increased our investment in what repeatedly worked.
This changes the meaning of strategy. Strategy is not merely a document predicting everything the company will do for the next three years. It is also a system for allocating more resources toward what the company is learning works.
An experiment that succeeds once is interesting. An experiment that succeeds across different customers, periods or channels begins to reveal a pattern. When the pattern continues, the company can invest more confidently in people, technology and distribution.
This is how a small experiment becomes an important business direction.
It also prevents the founder’s enthusiasm from becoming the only reason an idea receives resources. The company is still led by conviction, but capital follows evidence.
This is particularly important when building in Africa, where access to replacement capital can be difficult and expensive. A founder cannot assume that another funding round will arrive to compensate for several large, untested bets. The company must learn as much as possible from every naira, dollar and hour it spends.
Cheap experiments are therefore not evidence of small ambition. They are a way of protecting large ambition from unnecessary failure.
Speed matters, but so does the speed of learning
Founders are frequently told to move quickly. Speed matters because opportunities close, competitors advance and companies with limited runway cannot spend indefinitely considering each decision.
However, movement alone is not progress. A team can build rapidly in the wrong direction, run many campaigns without understanding acquisition and launch several features without improving retention.
The useful measure is not only execution speed but learning speed.
How quickly can the company turn an assumption into a test, the test into evidence and the evidence into a better decision? How quickly can it discover that an approach is wrong before that approach consumes a substantial amount of capital?
This is one reason small experiments are powerful. They shorten the distance between opinion and reality.
If the test succeeds, the company gains the confidence to move faster. If it fails, the company loses less and can try another approach. Either outcome creates progress because uncertainty has been reduced.
Do not become emotionally loyal to an experiment
Founders often become attached to ideas because the ideas originated with them. Once time, reputation and money have been invested, stopping can feel like an admission that the founder lacked judgment.
This emotional attachment corrupts the experiment. Instead of asking what the result teaches, the founder begins looking for reasons the result should not count. The wrong customers were selected, the timing was imperfect, the team did not execute properly or the market needs more education.
Any of these explanations may be correct, but they must produce another bounded test. They cannot become reasons to fund the same disappointing outcome indefinitely.
The founder should remain committed to the problem while staying flexible about the solution. An experiment is not your identity. Its purpose is to reveal something you did not know.
Failure at a small scale is often valuable because it prevents failure at a large scale. The cheapest time to discover that an assumption is wrong is before the company has hired a department, built complex infrastructure and announced the strategy to the entire market.
The discipline is to keep experimenting
One of my recurring mantras is to continue experimenting. If a company is no longer experimenting, it is difficult to believe that it is still making meaningful progress.
This does not mean changing everything constantly. The company should preserve what is working while continuing to question what could work better. Stable operations and controlled experiments should exist together.
Strong opinions give a startup the courage to attempt what other people may not yet believe. Cheap experiments give the startup the humility to discover whether its opinion survives contact with reality.
The combination is powerful: believe strongly, test intelligently, learn quickly and increase the investment only when the evidence earns it.
Do not wait for certainty before acting, because certainty rarely arrives first. But do not make every conviction an expensive commitment either.
Decide what you believe. Design the smallest credible test. Define the result that matters. Then allow the evidence to determine how much further you should go.
Strong opinions should not produce long arguments. They should produce cheap experiments.
Leave a comment