I once read a very good book called Heart, Smarts, Guts, and Luck, written by Anthony K. Tjan, Richard J. Harrington and Tsun-Yan Hsieh. The book examines the different qualities entrepreneurs draw upon when building businesses, and its title captures something important about decision-making: leaders do not always have the luxury of acting with complete information.
Sometimes the numbers are available, the market has been studied and the likely outcome can be modelled with reasonable confidence. At other times, a leader has to make a consequential decision with fragmented information, an unfamiliar market and a limited amount of time. Waiting for perfect data may feel responsible, but it can become another way of refusing to decide.
The answer is not to abandon data and rely entirely on instinct. It is to make the best decision possible from the evidence available while designing every action to produce better data for the next decision.
That is the principle I have come to believe: when you do not have enough data today, everything you do should help you acquire better data tomorrow.
Data is not equally available in every market
This challenge becomes obvious when people transfer strategies from one country to another. A highly experienced digital marketer may have worked in a market such as the United States, where there is extensive information about locations, consumer groups, incomes, interests and purchasing behaviour. Advertising platforms, commercial databases and public records can help a business define its audience with considerable precision.
The same person may enter Nigeria and discover that the information environment is different. Data can be fragmented, inconsistently maintained or unavailable at the level required for the campaign. Even when a dataset exists, it may not have enough historical depth or broad adoption to support the same type of analysis used elsewhere.
Nigeria launched a new National Digital Alphanumeric Postcode System on 1 October 2026. It gives each addressable building an 11-character, location-linked postcode and has the potential to improve logistics, customer verification, emergency response, financial services and market analysis. This is an important development, but a newly launched system cannot instantly provide businesses with years of behavioural and commercial data. Adoption, integration and the accumulation of useful history will take time.
This difference matters because a strategy built on granular location data in one country cannot simply be copied into a market where the underlying information is still developing. The marketer’s experience remains valuable, but the method has to change.
Leaders must understand the quality of the data environment in which they are operating. “We used this successfully elsewhere” is not sufficient evidence that it will work here.
Do not import conclusions when you can import questions
When local data is scarce, it can be tempting to use information from another country as a substitute. International data may provide context, reveal possibilities and help you form an initial hypothesis, but it should not be treated automatically as proof of local behaviour.
Different markets have different incomes, cultures, infrastructure, regulations, distribution systems and levels of trust. A product that succeeds in the United States may address a real need in Nigeria while requiring a completely different price, message and route to the customer. Even two neighbourhoods within the same city may behave differently for reasons that are not visible in a foreign dataset.
I would therefore be cautious about importing somebody else’s conclusion. What can be imported more safely are the questions behind the analysis. What customer problem was the business trying to solve? Which behaviour indicated demand? What variables predicted conversion? How did the company test price sensitivity? Why did one distribution channel outperform another?
Use those questions to investigate your own market. External evidence can help you decide where to look, but local evidence should determine what you eventually believe.
Begin with a smart hypothesis
When you lack sufficient data, you still need a starting point. That starting point is a hypothesis: a clear, testable explanation of what you believe and why.
Suppose you believe your target customers are high-net-worth individuals who are concentrated in a particular area. Do not present this as established truth simply because the area has expensive houses or a prestigious reputation. State it as a hypothesis. Then define what evidence would support or weaken it.
You may discover that the area contains mostly older residents while the product appeals to younger customers. You may find that many of the people who own property there spend most of their time elsewhere. You may learn that the professionals who fit your profile live across several locations but gather through particular workplaces, clubs, churches, schools or online communities.
The initial hypothesis may be wrong, but a testable hypothesis is still useful because it tells you what to investigate. An undefined assumption merely influences your decisions from the background. A written hypothesis can be examined and corrected.
A good hypothesis should answer several questions:
- Who do we believe the customer is?
- What problem do we believe this person has?
- How is the customer solving it today?
- Why would the customer choose our solution?
- Where can this customer be reached?
- What behaviour would show that our belief is correct?
- What result would cause us to reject or revise the hypothesis?
The last question is important. If no possible outcome can change your mind, you do not have a hypothesis; you have a belief that has been protected from evidence.
Create primary data through conversations
When large datasets are unavailable, talk to people. Interviews will not replace every form of quantitative research, but they can reveal language, motivations, fears and decision processes that numerical summaries may miss.
Start with a carefully chosen sample. Speak with existing customers, prospective customers and people who considered the product but chose an alternative. Talk to colleagues, friends and family when they genuinely fit the relevant profile, then ask them to introduce you to others whose circumstances differ from theirs. Friends of friends can help you move beyond the familiar opinions of your immediate circle.
The quality of the insight depends on the quality of the questions. Asking whether someone likes your idea is rarely useful because people often want to encourage you. Ask about actual behaviour instead. What did they do the last time the problem occurred? What did it cost them? Who was involved in the decision? Which alternatives did they consider? What made them hesitate? Have they ever paid to solve it? What would make the problem urgent?
People are often poor predictors of what they will do, but they can describe what they have already done. Past behaviour does not guarantee future behaviour, yet it is usually more reliable than a polite expression of interest.
Record the conversations systematically. If five team members conduct interviews without using a common structure, they may return with five collections of impressions that cannot be compared. Establish the key questions, capture responses consistently and identify repeated patterns. Qualitative information becomes more useful when it is organized rather than remembered selectively.
Use a sample, but understand what it can tell you
A small sample can help a business make an initial decision, especially when the alternative is guessing. However, the sample must reflect the question you are trying to answer.
If you interview only people in your professional network, you may learn how people like you think without learning how the broader market behaves. If you speak only with existing customers, you may understand satisfaction but miss why other people refuse to buy. If you test only in a wealthy neighbourhood, the result may say little about price sensitivity elsewhere.
You do not always need a massive research exercise before acting. You do need to know the limits of your sample. Treat the result as directional evidence, not universal truth. Then design the next test to include the people or conditions that the first sample omitted.
This is how evidence grows: one carefully interpreted layer at a time.
Turn execution into an instrument for learning
Research alone will eventually reach its limit. At some point, the business must execute because real behaviour under real conditions will answer questions that interviews cannot.
Launch the smallest test that can produce credible evidence. Run the campaign in a defined location. Offer the product to a particular customer group. Test two messages. Try one acquisition channel for a fixed period. Give the experiment a budget, an expected outcome and a stopping rule.
Before beginning, write down what you expect to happen. If we spend this amount, how many qualified enquiries should we receive? If this customer problem is urgent, how many people should complete the next step? If the price is acceptable, what conversion rate would make the economics work?
This protects the team from redefining success after seeing the result. Without an expectation, almost any outcome can be explained as encouraging. With an expectation, you can compare what you believed with what actually occurred.
When the experiment ends, do not ask only whether it succeeded. Ask what it taught you. Perhaps the target audience was correct but the message was weak. Perhaps people engaged with the advertisement but refused to pay. Perhaps the area produced leads, but customer acquisition cost made the channel unsustainable. Each result should improve the next hypothesis.
Execution should leave the business more informed than it was before, even when the immediate outcome is disappointing.
Build data collection into the operation
A company should not remain permanently dependent on intuition. Every campaign, product interaction, sales conversation and support issue should contribute to an increasingly useful body of information.
Decide what needs to be captured before launching an initiative. For marketing, this may include source, location, customer segment, message, cost, enquiry quality, conversion and revenue. For product, it may include activation, frequency of use, completion, failure points and retention. For sales, it may include industry, company size, decision maker, objections, sales-cycle length and reason for winning or losing.
Collecting everything is not the goal. Excessive measurement can produce a large amount of noise and create work nobody uses. Capture the information that helps the company choose what to do next.
The data must also be clean enough to compare. Agree on definitions. If one team calls every sign-up a lead while another counts only qualified prospects, the apparent performance cannot be interpreted properly. If customer locations are recorded in several inconsistent formats, geographic analysis will remain unreliable even when the database becomes large.
Scarcity of public data makes the discipline of building owned data even more valuable. Over time, a company that records its own transactions and customer behaviour consistently can develop an advantage that competitors cannot easily copy.
Combine heart, smarts and guts without confusing their roles
The title Heart, Smarts, Guts, and Luck is useful because decisions under uncertainty require more than one quality. Heart helps you care deeply enough about the problem and the people affected by it. Smarts helps you frame the question, examine evidence and understand the economics. Guts helps you act before every uncertainty has disappeared. Luck reminds you that outcomes are never entirely within your control.
The qualities should support one another. Guts without smarts can become recklessness. Smarts without guts can become endless analysis. Heart without evidence can make a leader defend an idea merely because it is personally meaningful. A leader needs enough self-awareness to recognize which tendency is dominating the decision.
Data does not remove the need for judgment. It improves the material on which judgment operates. Likewise, judgment does not excuse the absence of data. It helps you decide how to proceed responsibly while building the evidence you lack.
Make the decision, learn and adjust
When information is incomplete, begin with what is known, state what is assumed and identify what must be learned. Talk to the people closest to the problem. Build a sample. Form a smart hypothesis. Execute a controlled test. Measure the result, revise the idea and move again.
Do not wait indefinitely for a dataset that may never arrive, but do not use scarcity of data as permission to remain careless. Your first decision may rely heavily on experience, observation and informed judgment. Your second decision should benefit from what the first one taught you. If you are still making the tenth decision with no more evidence than you had at the beginning, the problem is no longer the market’s lack of data. It is your failure to learn.
The objective is not to eliminate uncertainty before acting. The objective is to create a disciplined cycle in which every action reduces uncertainty.
Start with the best judgment you can form. Execute carefully enough to learn. Then allow the evidence to make you smarter.
References
- Anthony K. Tjan, Richard J. Harrington and Tsun-Yan Hsieh, Heart, Smarts, Guts, and Luck: What It Takes to Be an Entrepreneur and Build a Great Business, Harvard Business Review Press, 2012.
- Federal Ministry of Communications, Innovation and Digital Economy, Federal Government Launches National Digital Alphanumeric Postcode System, 1 October 2026.
Leave a Reply