What Market Research Cannot Tell You: Making Decisions Under Uncertainty

Good market research can narrow uncertainty.

It cannot remove uncertainty from a business decision.

That distinction matters because research can become a form of delay: one more search, one more provider, one more dataset, and perhaps the correct answer will finally become obvious.

Often it will not.

The more useful question is whether the remaining uncertainty is sufficiently understood to make the next bounded decision.

Research Reduces Uncertainty; It Does Not Remove It

Public research can often establish observations such as:

• comparable services are publicly offered;
• defined customer groups exist;
• providers use particular service structures;
• some public price observations are available;
• geographic markets differ on selected characteristics.

It usually cannot establish:

• how many future customers will buy;
• whether one positioning message will outperform another;
• future revenue or profitability;
• future competitive actions;
• individual execution quality;
• whether a chosen approach will succeed.

These are not necessarily failures of research.

They are boundaries created by the question and the available evidence.

Five Common Sources of Uncertainty

1. Incomplete public information

Many providers do not publish:

• prices;
• customer counts;
• project volume;
• conversion rates;
• utilization;
• revenue.

Public provider research therefore observes only part of the market.

2. Non-comparable data

Two legitimate sources may define their subject differently.

Definitions can vary by:

• geography;
• industry;
• customer type;
• time period;
• service unit.

Combining them without accounting for those differences can produce a misleading conclusion.

3. Stale evidence

Market and economic conditions change.

A historical source may provide useful context without being current enough to support a present-tense claim.

When a present-tense factual claim depends on current conditions, verify it with a current source before publication. If current verification is unavailable, narrow the statement so it no longer depends on current evidence.

4. Proxy measurements

Researchers often use indirect indicators when direct demand cannot be observed.

Search interest, provider activity, customer counts, or broader industry trends may all be relevant.

None is identical to purchases of a specific service.

The proxy should remain identified as a proxy.

5. Future behavior

Some of the most important variables do not yet exist.

Future customers have not made future purchasing decisions.

Competitors have not made future pricing changes.

The professional has not accumulated future operating experience.

Historical research cannot turn those future events into known facts.

Use a Known / Uncertain / Unknowable Framework

Separate questions into three categories.

Known enough for the current decision

Available evidence is sufficiently clear and relevant.

Example:

The defined customer group can be identified in suitable public information.

Uncertain but researchable

Additional work could materially improve the answer.

Example:

More comparable public pricing observations may exist and could change the scenario comparison.

Not reasonably knowable in advance

Additional research is unlikely to produce a reliable answer.

Example:

Exactly how many customers will purchase this service next year?

This classification helps prevent unlimited research on questions that cannot be resolved.

Decision Thresholds Matter

A decision threshold defines how much evidence you require before taking the next step.

Suppose a professional is comparing three positioning alternatives.

The threshold might require:

• several credible comparable-provider observations for each option;
• fit with relevant prior experience;
• clearly definable service scope;
• understanding of material contradictory evidence;
• no reliance on a prediction of future customer volume.

Once those conditions are met, additional research may have declining value.

The threshold should be defined before endless searching begins.

When More Research Is Worth Doing

Additional research can be useful when:

• a missing fact could materially change the decision;
• two sources conflict and the reason may be resolvable;
• a price comparison rests on too few comparable observations;
• geographic data uses the wrong customer definition;
• an assumption is both high-impact and realistically researchable.

The question is not:

Can I find more information?

Usually you can.

The better question is:

Could additional information materially change the decision?

When It May Be Reasonable to Stop

Stopping research can be appropriate when:

• the decision is reversible;
• the consequences of an imperfect choice are bounded;
• core evidence is sufficient;
• remaining uncertainty primarily concerns future behavior;
• additional sources repeat information already collected;
• the research question has reached the practical limits of public evidence.

This does not make the decision safe or certain.

It makes the decision bounded.

Put Limitations in the Conclusion

A weak analysis hides limitations in a footnote.

A strong analysis makes them part of the actual conclusion.

For example:

Public evidence supports treating Markets A and B as meaningfully different on customer concentration and comparable-provider visibility. Pricing evidence remains too limited for a confident market-level comparison. The research can therefore support a geographic-prioritization decision, but not a conclusion about future customer demand or financial results.

The conclusion is narrower than a bold recommendation.

It is also clearer about what the evidence does and does not support.

Confidence Should Be Specific

Avoid assigning one broad “confidence level” to an entire market or concept.

Confidence should attach to particular observations.

For example:

Higher confidence: A clearly defined public measure from an appropriate official source.

Moderate confidence: A provider-positioning pattern observed across a defined public sample.

Lower confidence: A pricing comparison based on only a few partially comparable offers.

Confidence belongs to evidence quality, not to optimism.

Evidence, Assumptions, and the Final Decision

A decision-support record might say:

Evidence

• customer segment is clearly defined;
• comparable providers exist;
• geographic differences are observable.

Assumptions

• remote delivery reduces the importance of physical provider location;
• current delivery-time estimates remain reasonably stable.

Uncertainty

• future customer response;
• future competitive pricing;
• actual project volume.

Decision

Prioritize one market for further operating observation rather than assuming that national expansion is already justified.

This is a bounded conclusion.

Uncertainty Is Not the Same as Ignorance

After disciplined research, you may still be uncertain.

But you may now know:

• what the evidence supports;
• what contradicts the preferred interpretation;
• what information is missing;
• which assumptions matter most;
• what is unlikely to become knowable before action.

That is a meaningful improvement in decision quality.

Practical Conclusion

Market research does not need to eliminate uncertainty to be useful.

Its job is to improve the structure of the decision.

A strong research process identifies:

• what evidence supports;
• what remains assumed;
• what remains uncertain;
• what additional research could realistically resolve;
• what no reasonable research can promise in advance.

At that point, the decision-maker can determine whether the remaining uncertainty is acceptable for the next step.

Harborfield closing note

The Complete Solo Practice Research Dossier is built around this distinction: research can organize evidence, scenarios, assumptions, limitations, and decision criteria without pretending to remove future uncertainty. Its value lies in making the decision structure clearer, not in promising what the market will do next.