In this article · 7 parts
More research helps when it can change a decision you are about to make. It stops helping when you keep collecting general advice while the important uncertainty requires a different kind of evidence: an official answer, a customer conversation or experience of doing the work.
If you have spent months reading about starting a business alongside everything else you do, do not assume you lack courage. You may be trying to answer a real question with the wrong method. You may also lack the time, money or permissions to act. Those are constraints to address, not personality flaws.
The next useful step is to name the decision and identify what would allow you to make it.
1Replace the topic with a decision
“Research a mobile repair business” can expand indefinitely. There is always another tool, competitor, qualification or marketing channel to investigate.
“Decide whether I can offer one specific repair safely within my skills and available hours” has boundaries. It lets you distinguish questions that matter now from questions about a future business you may never operate.
Write the decision in ordinary language. Then complete these sentences:
- I currently believe that…
- The evidence for that belief is…
- I would change my decision if I found…
- The next commitment would expose me to…
The last sentence keeps the standard of evidence proportionate. Choosing which person to interview is different from accepting work on an expensive item, signing a lease or leaving employment. A reversible conversation does not need the same preparation as an obligation you cannot easily undo.
2Different uncertainties need different evidence
Use the following distinctions to choose your next action. They are a practical sorting method, not a diagnostic test.
| The uncertainty | Useful evidence | What will not settle it |
|---|---|---|
| Am I allowed and competent to do this work? | Relevant authority, insurer, professional requirements and contract review | A competitor doing something similar |
| Does this customer have a problem worth addressing? | Specific accounts of recent situations and existing alternatives | General enthusiasm for your idea |
| Will someone buy this defined offer? | A genuine buying decision under clear terms, when you can fulfil it properly | Survey answers about hypothetical spending |
| Can I deliver within my available time? | An appropriately bounded delivery test, including preparation and recovery | Another productivity article |
| Will the business work indefinitely? | No single source can establish this; use staged decisions and continuing review | A confident forecast with untested assumptions |
The uncertainty: Am I allowed and competent to do this work?
- Useful evidence
- Relevant authority, insurer, professional requirements and contract review
- What will not settle it
- A competitor doing something similar
The uncertainty: Does this customer have a problem worth addressing?
- Useful evidence
- Specific accounts of recent situations and existing alternatives
- What will not settle it
- General enthusiasm for your idea
The uncertainty: Will someone buy this defined offer?
- Useful evidence
- A genuine buying decision under clear terms, when you can fulfil it properly
- What will not settle it
- Survey answers about hypothetical spending
The uncertainty: Can I deliver within my available time?
- Useful evidence
- An appropriately bounded delivery test, including preparation and recovery
- What will not settle it
- Another productivity article
The uncertainty: Will the business work indefinitely?
- Useful evidence
- No single source can establish this; use staged decisions and continuing review
- What will not settle it
- A confident forecast with untested assumptions
Some questions require several methods. A customer saying they want a service does not establish that you can legally provide it. Permission to provide it does not establish demand.
3Set an end point before opening another search
A research task needs a question, a source standard and an end condition.
For a legal or regulatory question, the end condition might be an answer from the relevant authority or a qualified adviser who has seen your actual circumstances. An arbitrary reading deadline cannot make an unresolved safety issue disappear.
For a market question, the end condition might be enough evidence to choose the next low-exposure test. You do not need a complete industry report to discover that your likely customers require appointments during your work shifts.
Decide what you will do if the information remains unavailable. You might narrow the service, postpone taking orders or ask a more specific question. Otherwise, “I need more certainty” can become an instruction to keep searching forever.
4Illustrative example: the course that does not answer the question
Made up, to show the idea
A shift worker wants to offer clothing alterations. They have practised simple repairs but have not delivered paid work. They are considering another general business course because they still do not feel ready.
Their actual uncertainties are more specific. Can they reliably complete the alterations they intend to offer? How would fittings and collections work around changing shifts? What happens if they damage a customer's garment? Are potential customers looking for simple repairs or specialist work beyond their competence?
Another broad course may be interesting, but it does not directly resolve those questions. The reader instead limits the proposed offer, checks the relevant business and insurance arrangements, times practice on material they own and asks prospective customers about recent alteration purchases.
Suppose those conversations reveal that customers chiefly want rapid turnaround on valuable occasion wear. The reader cannot promise that around their current rota. This is useful evidence against the original offer. It does not prove there is no market for every alteration service.
They might investigate a narrower, slower service or choose a different idea. Buying the course first would not necessarily have been wasteful, but it would have delayed the decision that currently matters.
5Research can produce a decision to stop
A useful investigation does not have to end with a launch. Rejecting a poor fit before committing money or somebody else's time is a legitimate result.
Research on entrepreneurial decision-making has examined structured hypothesis testing rather than certainty-seeking. A 2024 replication involving four randomised trials found increased idea termination and more nuanced effects on strategic changes. That supports taking rejection seriously as an outcome; it does not prove that the framework in this article will improve your income or remove anxiety. The study populations and interventions are not the same as every reader's circumstances.
You can apply the underlying question without turning your life into a research project: what evidence would make this idea less attractive? If every possible result becomes a reason to continue, the test cannot protect you from a weak idea.
6Keep a small decision record
Use a note you can return to after a long shift. Record the question, the source or observation, what it suggests, what it cannot establish and the next decision.
For example: “Three people described difficulty finding evening collections. They were recruited through a friend, so they may not represent other buyers. I will investigate whether fixed collection windows solve the problem before offering a service.”
The number of conversations is an observation, not a magic validation threshold. A small, biased sample should support a small conclusion. Write down conflicting evidence too, especially if it came from somebody more like the intended customer than your enthusiastic friends.
This record also prevents repeated work. When the same worry returns, you can see whether anything has changed or whether you are reopening a decision without new information.
7If the next action still feels impossible
Check the action itself. Is it too large? Does it require an uninterrupted afternoon you never have? Are you assuming you can spend money that is already committed? Is another person's agreement missing?
Reduce the task while preserving its purpose. Drafting one precise question for an insurer is smaller than sorting out every business requirement. Asking one buyer about their most recent purchase is smaller than conducting a market survey. Neither promises a complete answer, but each can move a particular uncertainty forward.
If the action is possible but the outcome remains uncomfortable, leave room for that discomfort. A decision can be carefully prepared and still uncertain. This article is a planning method, not treatment for distress, and you do not need to turn personal difficulty into a business productivity target.
Before your next research session, write: “At the end of this, I want to decide whether to…”. If you cannot finish the sentence, choose the decision first. If you can, find the smallest credible source of evidence that could change your answer.
General information to help you think things through. Not regulated financial, legal or tax advice.