How to Conduct Market Research for a Solopreneur Business

Market research is the process of collecting and interpreting evidence about:

• Customers
• Problems
• Purchasing behavior
• Competitors
• Alternatives
• Prices
• Distribution
• Market size
• Market change

For a solopreneur, the purpose is practical.

Research should help you decide:

• Whether to pursue a market
• Which customers to prioritize
• Which problem to solve
• How to describe the offer
• What customers currently pay
• How they find solutions
• Whether enough demand exists
• Which risks require further testing

Market research cannot guarantee that a business will succeed.

It can prevent avoidable mistakes such as:

• Building for the wrong customer
• Solving a minor problem
• Entering a market you cannot reach
• Setting a price below delivery cost
• Misreading search interest as purchasing demand
• Underestimating established alternatives
• Using outdated or incomparable data

The strongest research combines several forms of evidence rather than relying on one survey, report, keyword tool, or conversation.

How to Conduct Market Research

Use this sequence:

1. Define the business decision.
2. Write the assumptions behind it.
3. Identify the people, businesses, or transactions being studied.
4. Separate what is already known from what must be researched.
5. Gather official and credible secondary data.
6. Map direct competitors and customer alternatives.
7. Interview potential customers.
8. Observe current behavior and workflows.
9. Use surveys only when structured quantitative data are needed.
10. Research prices and purchasing conditions.
11. Estimate the reachable market from the bottom up.
12. Investigate customer-acquisition channels.
13. compare evidence from different sources.
14. Record uncertainty and limitations.
15. Make a go, revise, test, or stop decision.

Market research should progress toward a commercial test.

It should not become a permanent substitute for offering something customers can buy.

Market Research at a Glance

What Market Research Can and Cannot Tell You

Market research can help you understand:

• Which customers experience a problem
• How customers currently respond
• What alternatives exist
• How customers describe the situation
• Where purchasing authority sits
• How much current solutions cost
• Whether the market is growing, shrinking, or changing
• Which channels customers use
• Which assumptions deserve a commercial test

Market research cannot prove:

• That customers will buy your future offer
• That your execution will be effective
• That market conditions will remain unchanged
• That a large market will be easy to enter
• That survey respondents will behave as predicted
• That your business will be profitable

Research supports decisions under uncertainty.

Actual purchases provide stronger evidence than stated interest.

Start With a Decision, Not a Topic

Weak research goal:

Research the market for AI services.

This is too broad to direct useful work.

Stronger research goal:

Decide whether to offer AI-assisted document-classification setup to independent accounting practices with fewer than 20 employees.

The second version makes it possible to identify:

• The customer population
• The current workflow
• The buyer
• The alternatives
• The likely price
• The regulatory and data risks
• The evidence needed to proceed

Write the decision in one sentence

Use:

I need to decide whether to [action] for [customer] in order to solve [problem].

Example:

I need to decide whether to build a paid monitoring service for small cosmetics brands that need to track changes in EU product requirements.

Set a research deadline

Research expands easily when no stopping point exists.

Choose a deadline appropriate to the decision:

• Seven days for an initial opportunity scan
• Two to four weeks for a first offer
• Longer for regulated products, manufacturing, or major capital commitments

The deadline does not mean all uncertainty must disappear.

It means you must decide whether the existing evidence justifies the next investment.

Step 1: Write a Market Research Brief

Before collecting data, create a short research brief.

Research decision

What must be decided?

Customer hypothesis

Which customer is being considered?

Problem hypothesis

Which condition or consequence is believed to exist?

Offer hypothesis

What might the business provide?

Geographic scope

Which country, region, city, or language market applies?

Evidence needed

What evidence would increase confidence?

Disconfirming evidence

What finding would make you stop or revise the idea?

Constraints

Record limitations involving:

• Budget
• Time
• Language
• Customer access
• Data availability
• Qualifications
• Regulation

Decision date

When will the research be reviewed?

Example research brief

A research brief prevents unrelated information from dominating the process.

Step 2: Convert Assumptions Into Research Questions

Every business idea contains assumptions.

Common assumptions include:

• The customer exists in sufficient numbers.
• The problem occurs frequently.
• The customer considers it important.
• The buyer has a budget.
• Existing alternatives are inadequate.
• The market can be reached economically.
• The proposed solution can be delivered profitably.

Write each assumption as a question.

Prioritize questions according to risk.

Research the assumption most capable of invalidating the business first.

Do not begin by answering the easiest questions.

Step 3: Define the Research Population

Before gathering data, specify exactly what is being studied.

Possible units include:

• Individuals
• Households
• Businesses
• Business locations
• Transactions
• Products
• Website searches
• Software accounts

These units are not interchangeable.

One company may operate:

• Several locations
• Several brands
• Several websites
• Multiple legal entities

One person may own several businesses.

Define inclusion criteria

For a B2B market, criteria might include:

• Industry
• Number of employees
• Revenue range
• Location
• Software used
• Business model
• Customer type
• Current purchasing behavior

For a consumer market, criteria might include:

• Age
• Household type
• Location
• Income
• Purchasing event
• Product ownership
• Recent behavior

Only include a characteristic when it affects:

• The problem
• The purchasing decision
• The offer
• The price
• The delivery method

Define exclusions

Explicitly exclude groups that look similar but do not belong in the research.

For example:

Include independent accounting firms using cloud document portals. Exclude internal corporate accounting departments and firms using only paper records.

Clear boundaries improve:

• Search queries
• Interview recruitment
• Market estimates
• Competitor comparisons

Primary and Secondary Market Research

Market research uses two broad forms of evidence.

Secondary research

Secondary research examines information collected by someone else.

Sources include:

• Official statistics
• Company registers
• Industry reports
• Financial statements
• Academic research
• Competitor websites
• Product reviews
• Search data
• Trade publications
• Government regulations

Secondary research is useful for understanding:

• Market structure
• Customer counts
• Geography
• Broad spending
• Industry changes
• Competition
• Regulation

Its main weakness is that the available categories may not match your exact customer or problem.

Primary research

Primary research collects new evidence directly for your decision.

Methods include:

• Customer interviews
• Surveys
• Observation
• Diary studies
• Prototype tests
• Price tests
• Sales conversations
• Paid pilots

Primary research is useful for understanding:

• Customer language
• Current behavior
• Buying triggers
• Internal processes
• Objections
• Unmet needs
• Willingness to commit

Its main weakness is that small or poorly selected samples can produce misleading conclusions.

Use both

Secondary research may show that 10,000 relevant businesses exist.

Primary research may reveal that most do not recognize the problem or already solve it satisfactorily.

Primary research may reveal strong interest among ten customers.

Secondary research may show that the total reachable population is too small for the intended pricing model.

The forms of evidence should test one another.

Qualitative and Quantitative Research

Qualitative research

Qualitative research explores:

• Why something happens
• How customers think
• How work is performed
• Which language customers use
• Why alternatives are accepted or rejected

Common methods include:

• Interviews
• Observation
• Open-ended responses
• Review analysis

Qualitative findings are detailed but usually should not be presented as population percentages.

Quantitative research

Quantitative research measures:

• How many
• How often
• How much
• Which option
• How results differ between groups

Common methods include:

• Surveys
• Transaction analysis
• Official statistics
• Search and traffic data
• Pricing data

Quantitative evidence can estimate scale, but its reliability depends on:

• Definitions
• Sampling
• Measurement
• Response quality
• Data freshness

Recommended sequence

For a new solopreneur opportunity:

1. Use qualitative research to discover relevant questions.
2. Use quantitative evidence to estimate frequency or scale.
3. Use a commercial offer to test behavior.

Beginning with a survey can force customers into answer choices based on assumptions you have not yet tested.

Step 4: Conduct Secondary Market Research

Begin with broad evidence, then narrow toward the exact customer.

Official statistical sources

Official sources can help estimate:

• Population
• Households
• Businesses
• Employment
• Income
• Consumer expenditure
• Industry sales
• Business creation
• Geographic distribution

The U.S. SBA framework recommends examining demand, market size, economic indicators, customer location, market saturation, and prices paid for alternatives.

U.S. market data

The current Census Builder, version 5.11 as of March 2026, combines selected demographic and economic data for entrepreneurs and small-business researchers. It can be used to examine local customer characteristics, industry activity, and potential competition.

The 2022 Economic Census remains the official five-year measure of U.S. employer businesses, providing national, state, and local statistics by industry and geography. Its reference year must be distinguished from its later publication date.

The BEA data provide monthly, quarterly, and annual estimates of U.S. consumer spending. Broad spending growth should be used as economic context rather than evidence that consumers will purchase one specific product.

European market data

Eurostat recorded 33.1 million active enterprises in the EU business economy in 2023, employing 162.2 million people. SMEs represented 99.8% of those enterprises. These Eurostat figures show the scale of the broad business population, but a useful market estimate must still isolate the relevant industry, size, geography, technology, and purchasing situation.

Eurostat’s business demography data can be used to study active enterprises, business births, deaths, and survival across covered industries.

For consumer markets, the household survey provides expenditure data by consumption category and selected household characteristics. Eurostat explicitly identifies market research as one possible use of these statistics.

National statistics offices

National statistical offices may provide more detailed information about:

• Local population
• Household spending
• Business registrations
• Employment
• Income
• Tourism
• Housing
• Trade

Use the national source when it offers more precise or recent information than an international database.

Company registers

Official company registers can help identify:

• Active companies
• Incorporation dates
• Registered activities
• Directors
• Accounts where publicly available
• Geographic concentration

The UK Companies House register can be searched free of charge and includes advanced search functions. Its guidance was updated in June 2026. Registered company information can support competitor identification and market counting, but it does not prove that every registered company actively trades in the expected market.

Company-register data may omit:

• Sole traders
• Informal activity
• Closed locations
• Actual customer segments
• Current revenue by product

Treat a register as one source rather than a complete market map.

How to Evaluate a Secondary Source

For each source, record:

Reference date vs. publication date

A page updated in 2026 may still report data from 2022 or 2023.

Always record both:

• The publication or update date
• The period the statistics describe

Recent publication does not automatically mean recent underlying data.

Check definitions before comparing figures

Two sources may define “small business” differently.

They may use:

• Employee thresholds
• Revenue thresholds
• Legal status
• Industry exclusions
• Employer-only data
• Employer and nonemployer data

Do not combine figures until you understand the definitions.

Step 5: Research the Competitive Environment

Competitor research should explain how customers currently solve the problem.

It should not become a collection of screenshots or an exercise in copying features.

Identify four forms of competition

Direct competitors

They sell a similar solution to the same customer.

Indirect competitors

They solve the same problem through a different method.

Internal alternatives

The customer uses employees, spreadsheets, or an internal process.

Inaction

The customer tolerates the problem.

Inaction can be the strongest competitor when:

• The consequence is unclear
• Switching is difficult
• The problem is infrequent
• The current process is considered good enough

Build a competitor list

Use:

• Search engines
• Marketplaces
• Industry directories
• Customer interviews
• Software directories
• Trade events
• Company registers
• Product reviews
• Referral partners

Do not include a company only because it uses similar language.

Confirm that it serves the same:

• Customer
• Problem
• Geography
• Price range
• Purchasing situation

Competitor research matrix

Separate fact from interpretation

Fact:

The competitor charges €99 per month and includes five user accounts.

Interpretation:

The competitor appears to target small teams rather than solo professionals.

Record these separately.

An interpretation should not be presented as confirmed fact.

Read customer reviews carefully

Reviews may reveal:

• Repeated failures
• Missing functions
• Support expectations
• Switching reasons
• Customer language
• Features that attract buyers

Reviews are subject to selection bias.

Customers who leave reviews may differ from the majority of users.

Look for repeated patterns across:

• Positive reviews
• Negative reviews
• Several platforms
• Direct customer interviews

Do not copy the competitor

Competitor research should help answer:

• Which expectations are already established?
• Which customers remain poorly served?
• Which promises are difficult to fulfil?
• Which delivery models appear viable?
• What must a new business explain differently?

A competitor’s visible website does not reveal:

• Profitability
• Customer-acquisition cost
• Refunds
• Owner workload
• Customer concentration
• Internal failures

Step 6: Conduct Customer Interviews

Customer interviews help you understand behavior that is not visible in public data.

Interview people who:

• Match the intended customer
• Have experienced the problem
• Influence or control the purchase
• Recently chose a solution
• Recently rejected a solution

Do not interview only:

• Friends
• Supportive followers
• Other entrepreneurs
• People who enjoy discussing ideas
• Customers outside the intended market

How many interviews do you need?

There is no universal number.

For an early market investigation, ten to fifteen relevant interviews may reveal recurring themes.

The quality of the sample matters more than reaching an arbitrary total.

Fifteen conversations with buyers who recently experienced the problem may be more informative than 100 conversations with a general audience.

Do not present interview findings as population percentages unless the research design supports that use.

Customer interview guide

Context

• What is your role?
• Which part of this process do you control?
• Who else is involved?

Recent behavior

• When did this problem last occur?
• What happened?
• What caused it?
• What did you do next?

Consequence

• How much time did it require?
• What was delayed?
• Did it create a financial cost?
• Who was affected?

Existing solution

• Which tools or providers did you use?
• What worked?
• What remained difficult?
• What did the solution cost?

Buying process

• Who chose the solution?
• Who approved the expense?
• Which alternatives were considered?
• What prevented a purchase?

Trigger

• What made the issue important at that moment?
• When would you address it again?

Success

• How would you know the problem was solved?
• Which result matters most?

Ask about specific past events

Weak:

Would you pay for an automated reporting tool?

Stronger:

How did you prepare the last monthly report?

Weak:

Is cybersecurity important to your company?

Stronger:

What happened the last time an employee or contractor needed access to a business account?

Past behavior reveals:

• Actual process
• Existing spending
• Real constraints
• Purchasing authority

Hypothetical answers may reflect politeness or optimism rather than likely behavior.

Do not pitch too early

When you explain your solution at the start, the interview can become:

• A sales conversation
• A request for approval
• A feature brainstorm
• A debate about your idea

Understand the customer’s current reality before introducing the proposed offer.

Record exact language

Customer wording can improve:

• Positioning
• Website copy
• Sales questions
• Search research
• Product instructions

Do not edit the notes immediately into your preferred terminology.

Preserve the original meaning.

Step 7: Observe the Existing Workflow

Customers do not always describe work accurately.

They may:

• Forget steps
• Combine several activities
• Underestimate time
• Omit exceptions
• Describe the official process rather than the real process

Observation can reveal:

• Repeated switching between tools
• Manual checking
• Delays
• Unofficial workarounds
• Hidden approvals
• Duplicate records
• Error-correction work

Observation methods

Depending on the business and privacy requirements, you might:

• Watch a screen-sharing session
• Ask the customer to demonstrate the process
• Review anonymized documents
• Examine support requests
• Map a recent transaction
• Ask the customer to keep a short activity diary

Create a process map

For each step, record:

• Person responsible
• Input
• Action
• Tool
• Waiting time
• Error risk
• Output

Example:

This may reveal that the apparent “reporting problem” is actually a product-data problem.

Step 8: Use Surveys for the Right Questions

A survey is useful when you already understand the topic well enough to ask structured questions.

Surveys can help estimate:

• Frequency
• Prevalence
• Preferences
• Current tools
• Spending ranges
• Differences between customer groups

A survey is less useful for discovering an unfamiliar problem because your answer choices may omit the most important behavior.

Define the survey population

Specify:

• Who qualifies
• How they will be reached
• Which geography applies
• Whether the respondent is the buyer, user, or both

A survey distributed to your followers measures your followers.

It does not automatically represent:

• All customers
• An industry
• A country
• A demographic group

Sampling methods

Probability sample

Each eligible unit has a known chance of selection.

This supports stronger population estimates but is usually expensive and difficult for a solopreneur.

Convenience sample

Respondents are selected because they are easy to reach.

Examples include:

• Newsletter subscribers
• Social followers
• Community members
• Existing customers

Convenience samples can provide useful directional evidence.

They should not be presented as statistically representative of the full market.

Customer sample

Existing customers can reveal:

• Satisfaction
• Usage
• Repeat-purchase behavior
• Product gaps

They cannot represent people who never discovered or purchased the offer.

Sample size

For a large population, a simple random sample of approximately 385 respondents produces a theoretical margin of error of about ±5 percentage points at the 95% confidence level when the measured proportion is near 50%.

This approximation assumes:

• Random selection
• Independent responses
• Accurate measurement
• No meaningful nonresponse bias

A large convenience sample does not become representative merely because it contains more responses.

Write neutral questions

Weak:

How useful would our affordable and convenient service be?

This wording introduces positive assumptions.

Stronger:

Which method did you use the last time you completed this task?

Ask one thing at a time

Weak:

How satisfied are you with the software’s price and customer support?

A customer may be satisfied with one and dissatisfied with the other.

Separate the questions.

Use specific time periods

Weak:

How often do you experience payment delays?

Stronger:

During the past three months, how many customer invoices were paid after the agreed deadline?

Make answer choices complete

Include:

• Relevant ranges
• “None”
• “Not applicable”
• “Other”
• “Do not know,” where appropriate

Consider question order

Earlier questions can affect the context respondents use when answering later questions. The survey guidance from the UK Government Analysis Function recommends planning questionnaire structure, testing questions, and considering how respondents understand and answer them.

Pretest the survey

Ask several people from the intended population to complete it while explaining:

• What they think each question means
• Whether answer choices fit
• Which terms are unclear
• Which information they cannot recall

The U.S. Census Bureau’s testing guidance notes that combining several pretesting methods is generally more effective than relying on only one method.

Step 9: Research Search Demand Properly

Search data can reveal how people express demand.

It can help identify:

• Common wording
• Product categories
• Comparisons
• Problems
• Purchase questions
• Seasonal interest
• Geographic differences

Search demand does not represent the entire market.

Customers may find solutions through:

• Referrals
• Direct sales
• Procurement
• Marketplaces
• Industry relationships
• Local visibility
• Existing software

Separate query types

Problem-aware searches

How to reduce product returns

Solution-aware searches

Product sizing software

Comparison searches

Product A vs Product B

Transactional searches

Product data audit price

Navigational searches

Company name login

The commercial value of a query depends on:

• Intent
• Customer value
• Competition
• Conversion
• Your ability to rank or advertise

High search volume does not automatically mean high commercial value.

Use Google Trends carefully

Google Trends reports relative interest rather than absolute search volume. Its data are normalized according to the location and time period selected, with results commonly shown on a scale from 0 to 100. The official Google guidance explains that the data are normalized to make terms easier to compare.

A value of 100 means the highest relative interest within that selected comparison.

It does not mean:

• 100 searches
• 100% market share
• A fixed amount of demand

Search volume is not customer count

One person may perform many searches.

Several searches may refer to the same purchase.

Some searches have no commercial intent.

Use search data to understand language and direction, then validate purchasing behavior through:

• Sales
• Leads
• Interviews
• Conversion tests

Step 10: Conduct Pricing Research

Pricing research should identify:

• What customers currently pay
• What the alternatives cost
• Which budget funds the purchase
• Which price structure customers understand
• Whether your price can support delivery

Sources of pricing evidence

Use:

• Public price pages
• Marketplace listings
• Supplier catalogues
• Customer invoices where lawfully shared
• Procurement records
• Competitor quotations
• Customer interviews
• Your own price tests

Compare complete costs

A competitor’s monthly fee may exclude:

• Setup
• Migration
• Support
• Usage fees
• Additional users
• Training
• Required integrations

Compare the total cost for the same customer situation.

Ask what customers paid previously

Useful questions include:

• What did the current solution cost?
• Was there an implementation fee?
• Which budget paid for it?
• Who approved it?
• Which price changes would require new approval?

Do not rely on direct willingness-to-pay questions

A customer may say they would pay €500 and later refuse a €200 offer.

Stated willingness may be influenced by:

• Politeness
• Lack of context
• Hypothetical budgets
• Desire to encourage you

A real priced offer is stronger evidence.

Use a price ladder

Present several realistic configurations:

Observe which trade-offs customers make.

Do not create arbitrary options only to push buyers toward the middle price.

Test price with a sale

The most reliable early pricing question is:

Will a suitable customer purchase this defined result under these conditions?

Record:

• Price
• Scope
• Customer
• Outcome
• Objection
• Purchase decision

Step 11: Estimate Market Size

Market size should be connected to:

• Customer count
• Price
• Purchase frequency
• Retention
• Owner capacity

A dramatic global market estimate is rarely useful to a new solopreneur.

Total addressable market

The full theoretical market for the category.

TAM = Total potential customers × Annual spending per customer

Serviceable available market

The portion you can serve given:

• Geography
• Language
• Product
• Qualifications
• Delivery model

SAM = Suitable customers × Annual spending per customer

Serviceable obtainable market

The portion you could realistically acquire given:

• Distribution
• Competition
• Credibility
• Capacity

SOM = Realistically obtainable customers × Annual revenue per customer

Use a bottom-up estimate

Suppose your research finds:

• 3,000 businesses in the broad category
• 800 use the relevant software
• 300 experience the specific problem
• 150 can be reached through available channels
• You can serve 20 per year
• Average annual revenue per customer is €5,000

Your capacity-limited obtainable revenue is:

20 × €5,000 = €100,000

The broad population matters less than the number of customers you can realistically acquire and serve.

Build low, base, and high cases

These are scenarios, not forecasts.

Each assumption should have supporting evidence.

Do not multiply incompatible data

Avoid calculations such as:

Global industry value × assumed 1% market share

unless you can explain how that share would be obtained.

A one-percent assumption often hides the hardest parts:

• Customer access
• Competition
• Sales cycle
• Capacity
• Trust

Step 12: Research Customer-Acquisition Channels

A market is not commercially accessible merely because customers exist.

Research where buyers:

• Search for information
• Compare providers
• Request recommendations
• Attend events
• Purchase solutions
• Ask for help

Channel research questions

• Can potential customers be identified individually?
• Which terms do they search?
• Which publications do they read?
• Which associations do they join?
• Which providers already have their trust?
• Are marketplaces involved?
• Is purchasing formal or informal?
• How long is the decision process?
• Does the buyer require a referral?

Channel evidence

Stronger evidence includes:

• Customers naming the same source
• Existing providers using the channel repeatedly
• Qualified leads from a small test
• Published procurement activity
• Search terms with relevant intent
• Referral partners willing to introduce customers

Weaker evidence includes:

• The channel has many users.
• Competitors have followers.
• A post received engagement.
• Advertising is available.

Test the channel manually

Before building a complete marketing system, perform a small test.

Examples include:

• Contact 30 qualified prospects.
• Publish one high-intent search page.
• Ask five complementary providers about referrals.
• Create one marketplace listing.
• Run a limited advertisement.

Track:

• Responses
• Qualified conversations
• Offers
• Sales
• Cost
• Owner time

A channel is useful when it produces suitable customers at workable economics.

Step 13: Research B2B Markets Differently

B2B research often involves several people and a longer buying process.

Identify:

• User
• Manager
• Technical reviewer
• Financial approver
• Legal or security reviewer
• Final decision-maker

Research the buying organization

Ask:

• Which department owns the problem?
• Which budget is used?
• Does the purchase require a contract?
• Are suppliers screened?
• Is insurance required?
• How long does approval take?
• Does the buyer need several quotations?
• Which system must the solution integrate with?

Count organizations rather than people

For a B2B service, the number of employees in the target profession may be less relevant than the number of organizations that can purchase.

For example:

• 20,000 professionals may work inside 1,500 firms.
• Only 300 firms may fit the intended size.
• Only 100 may use the relevant system.

Use the correct purchasing unit.

Investigate budget timing

Some B2B purchases occur:

• Monthly
• At renewal
• During annual planning
• After a trigger event
• Before an audit
• At the end of a financial year

Timing can determine whether a market appears uninterested or simply unavailable at that moment.

Step 14: Research Consumer Markets Differently

Consumer research should consider:

• Buyer
• User
• Household
• Purchasing occasion
• Frequency
• Income
• Location
• Substitutes
• Trust
• Convenience

The person paying may differ from the user.

Examples include:

• Parents buying for children
• Adult children buying for older parents
• Employers buying benefits for workers
• Gift buyers purchasing for recipients

Study purchasing occasions

Consumer behavior often changes according to:

• Season
• Life event
• Urgency
• Replacement cycle
• Promotion
• Location

A product may have modest average demand but strong demand during a specific event.

Examine household expenditure data carefully

Household budget data can show category spending.

It cannot reveal automatically:

• Brand demand
• Unmet needs
• Purchase intent
• Your likely market share

Use expenditure data to understand scale and distribution, then research the specific purchasing decision.

Step 15: Evaluate Data Quality

Not all data deserve equal weight.

Eurostat evaluates statistical quality using dimensions including relevance, accuracy, timeliness, accessibility and clarity, comparability, and coherence. These quality criteria provide a useful framework for evaluating market evidence.

Relevance

Does the source answer your actual question?

A national industry total may be accurate but irrelevant to a local specialist offer.

Accuracy

How closely does the figure represent the real condition?

Consider:

• Sampling
• Measurement
• Missing data
• Estimation
• Revision

Timeliness

Is the reference period recent enough?

The acceptable age depends on the market.

Population changes slowly.

Software prices, regulations, and platform usage may change quickly.

Accessibility and clarity

Can you understand:

• Definitions
• Units
• Method
• Limitations

A precise number without clear metadata may be difficult to use responsibly.

Comparability

Can results be compared across:

• Years
• Countries
• Sources
• Customer groups

Changes in classifications or survey questions can break comparability.

Coherence

Do related sources tell a compatible story?

Contradiction is not always an error.

Different sources may measure:

• Different populations
• Different time periods
• Different behaviors

Create a Source Log

Maintain a source log during research.

A source log makes later updates easier and reduces the risk of presenting stale data as current.

Step 16: Protect Research Participants and Data

Market research may involve personal or commercially sensitive information.

Collect only what is necessary.

Potentially sensitive material includes:

• Names
• Email addresses
• Job roles
• Recorded conversations
• Customer lists
• Internal workflows
• Revenue
• Screenshots
• Account information

For organizations subject to the GDPR, personal data must be processed lawfully and transparently, collected for specified purposes, limited to what is necessary, kept accurate, protected, and retained no longer than needed. The GDPR principles apply according to the circumstances of the processing.

Practical safeguards

• Explain why the information is being collected.
• Obtain appropriate permission before recording.
• Remove unnecessary identifiers from notes.
• Limit access.
• Use secure storage.
• Define a deletion date.
• Do not publish quotations without permission.
• Do not share confidential business information with unrelated tools.

Legal requirements vary by jurisdiction, data type, participant, and research purpose.

Step 17: Use AI Carefully in Market Research

AI can help with:

• Generating search terms
• Organizing source lists
• Transcribing interviews
• Grouping notes
• Comparing competitor information
• Summarizing documents
• Drafting survey questions
• Translating responses

AI should not be treated as the original source of market evidence.

Verify every factual claim

Confirm claims against:

• Official data
• Original documents
• Direct customer evidence
• Published competitor information

NIST’s AI framework provides a risk-management profile for generative AI and emphasizes evaluating and managing the trustworthiness of AI systems and outputs.

Protect confidential information

Do not upload:

• Identifiable interview transcripts
• Customer data
• Confidential contracts
• Unpublished financial information
• Proprietary product plans

to an external AI system without understanding:

• Data retention
• Training use
• Access
• Security
• Contractual restrictions

Use AI for coding, not final judgment

AI can suggest themes from interview notes.

The researcher should still:

• Read the evidence
• Check whether themes are supported
• Preserve contradictory responses
• Distinguish frequency from importance
• Avoid inventing consensus

Ten similar AI-generated summaries do not create ten independent sources.

Step 18: Analyze the Evidence

Do not summarize findings source by source.

Organize them around the decision.

Create an evidence matrix

Separate four levels of statement

Observation

Seven interviewees used spreadsheets.

Pattern

Spreadsheets were the most common workaround in the interview sample.

Interpretation

Customers may value flexibility more than automation.

Decision

The initial offer should improve the spreadsheet workflow before proposing replacement software.

Do not present an interpretation as though the customer stated it directly.

Look for contradictions

Contradictory evidence may indicate:

• Different customer segments
• Different problem severity
• Different company maturity
• Different purchasing authority
• A weak original definition

Do not remove inconvenient responses merely because they complicate the conclusion.

Step 19: Judge the Strength of Evidence

Not all evidence should receive equal weight.

Stronger evidence

• Recent customer purchases
• Signed agreements
• Paid pilots
• Transaction data
• Repeated observed behavior
• Official statistics with matching definitions
• Several independent sources reaching similar conclusions

Moderate evidence

• Customer interviews
• Existing workarounds
• Published competitor prices
• Qualified leads
• Search trends
• Industry reports with clear methodology

Weaker evidence

• Stated future intentions
• Social-media engagement
• Unsourced market estimates
• Friends’ opinions
• AI-generated summaries
• One viral example
• Large top-down market figures

A weak signal may justify further research.

It should not carry the same weight as a purchase.

Step 20: Set Decision Rules

Define the conditions for proceeding before interpreting the final evidence.

Example proceed criteria

• At least ten suitable customers confirm recent occurrences.
• Several customers already spend money or significant time.
• A clear buyer controls the expense.
• The reachable market exceeds the required customer count.
• The likely price supports delivery.
• At least two customers accept a paid pilot.

Example revise criteria

• The problem exists, but another customer owns the budget.
• Interest is strong, but the offer is too broad.
• The market is reachable, but the delivery model is too expensive.
• A related problem receives more urgency.

Example stop criteria

• Customers rarely experience the problem.
• Existing solutions are considered adequate.
• No buyer controls a suitable budget.
• Customer acquisition is impractical.
• The market cannot support the required economics.
• The work requires qualifications or staffing you do not possess.

Decision rules reduce the temptation to interpret every result as support for the original idea.

The Market Research Report

The final report should be short enough to inform action.

1. Decision

State what was being evaluated.

2. Customer definition

Define the included and excluded customers.

3. Problem

Describe the situation, consequence, frequency, and trigger.

4. Market evidence

Summarize:

• Customer counts
• Geography
• Industry conditions
• Relevant trends

5. Customer evidence

Include:

• Current behavior
• Workarounds
• Existing spending
• Buying process
• Exact language

6. Competitive environment

Explain:

• Direct competitors
• Indirect alternatives
• Inaction
• Price range

7. Distribution

State how potential customers can be reached.

8. Economics

Estimate:

• Price
• Purchase frequency
• Market size
• Delivery cost
• Required customer count

9. Risks

List assumptions still unresolved.

10. Recommendation

Choose:

• Proceed to paid test
• Revise the customer
• Revise the problem
• Revise the offer
• Continue research
• Stop

A 14-Day Market Research Plan

Day 1: Define the decision

Create the research brief.

Day 2: List assumptions

Rank them by risk.

Days 3–4: Gather secondary evidence

Research:

• Customer counts
• Industry structure
• Geography
• Regulation
• Broad spending

Days 5–6: Map competitors

Record:

• Customer
• Offer
• Price
• Distribution
• Proof
• Weaknesses

Days 7–10: Conduct interviews

Speak with relevant:

• Users
• Buyers
• Recent purchasers
• Former customers

Day 11: Map the current workflow

Identify:

• Workarounds
• Delays
• Costs
• Decision points

Day 12: Estimate market size

Build low, base, and high cases.

Day 13: Analyze evidence

Create:

• Evidence matrix
• Risk list
• Research limitations

Day 14: Make the decision

Choose the next commercial test.

A 30-Day Market Research Plan

Days 1–5: Define the market

Complete:

• Research brief
• Population definition
• Assumptions
• Evidence requirements

Days 6–10: Conduct desk research

Use:

• Official statistics
• Company registers
• Industry sources
• Search data
• Regulations

Days 11–15: Research competitors

Map:

• Offers
• Prices
• Proof
• Channels
• Customer reviews

Days 16–22: Conduct primary research

Complete:

• Customer interviews
• Workflow observations
• Relevant survey where appropriate

Days 23–25: Research pricing and channels

Test:

• Published pricing
• Realistic price options
• Prospect response
• Referral possibilities

Days 26–27: Estimate economics

Calculate:

• Reachable customers
• Required customer count
• Gross profit
• Owner capacity

Days 28–29: Synthesize findings

Separate:

• Facts
• Patterns
• Interpretations
• Unknowns

Day 30: Decide

Move to:

• Paid pilot
• Offer revision
• Additional targeted research
• Another market
• No further investment

Market Research Checklist

Research design

• [ ] A specific decision has been defined.
• [ ] The customer population is clear.
• [ ] Assumptions have been written.
• [ ] Disconfirming evidence has been defined.
• [ ] A deadline has been set.

Secondary research

• [ ] Official statistics have been checked.
• [ ] Reference periods are recorded.
• [ ] Definitions are understood.
• [ ] Company or customer counts have been estimated.
• [ ] Regulations have been reviewed.
• [ ] Data limitations are documented.

Competitive research

• [ ] Direct competitors are listed.
• [ ] Indirect alternatives are listed.
• [ ] Internal workarounds are understood.
• [ ] Published prices are recorded.
• [ ] Reviews and customer complaints have been examined.
• [ ] Facts are separated from interpretations.

Customer research

• [ ] Relevant customers have been recruited.
• [ ] Recent behavior was discussed.
• [ ] Existing spending was investigated.
• [ ] The buyer was identified.
• [ ] Buying triggers were recorded.
• [ ] Exact customer language was preserved.

Quantitative research

• [ ] The survey population is defined.
• [ ] Sampling limitations are understood.
• [ ] Questions are neutral.
• [ ] Questions ask one thing at a time.
• [ ] The questionnaire has been pretested.
• [ ] Results are not overstated.

Economics

• [ ] Reachable market size has been estimated.
• [ ] Current prices have been researched.
• [ ] Required customer count has been calculated.
• [ ] Owner capacity has been considered.
• [ ] Low, base, and high cases have been prepared.

Decision

• [ ] Supporting evidence is documented.
• [ ] Contradictory evidence is documented.
• [ ] Remaining uncertainties are clear.
• [ ] A proceed, revise, test, or stop decision has been made.

Common Market Research Mistakes

Researching a broad topic

“Research ecommerce” produces information without a clear decision.

Using only secondary data

Market statistics cannot explain the exact customer workflow or buying process.

Using only customer interviews

Several conversations cannot reliably establish the size of an entire market.

Interviewing only supportive people

Friends and followers may want to encourage you.

Pitching during research

The conversation becomes feedback on your solution rather than evidence about the current problem.

Asking only hypothetical questions

Stated future behavior is weaker than past action and payment.

Treating search volume as total demand

Many markets operate through referrals, procurement, marketplaces, and direct sales.

Treating a market report as validation

A large global category may contain no reachable customers for your offer.

Counting registrations as active customers

Company registers may contain inactive, dormant, unsuitable, or non-trading entities.

Comparing incompatible statistics

Sources may use different:

• Definitions
• Years
• Geographies
• Industries
• Business-size thresholds

Ignoring the reference period

Recently published data may describe activity from several years earlier.

Surveying before understanding the problem

Poor answer choices formalize the researcher’s assumptions.

Reporting percentages from a small interview sample

Qualitative interviews identify patterns and explanations rather than population prevalence.

Looking only for confirmation

Research should be capable of showing that the idea should not proceed.

Continuing research after the next test is obvious

When the main uncertainty is willingness to pay, further reading cannot replace a priced offer.

Frequently Asked Questions

What is market research?

Market research is the collection and interpretation of evidence about customers, demand, competitors, alternatives, prices, distribution, and market conditions for a business decision.

Why is market research important for a solopreneur?

A solopreneur has limited time and capital. Research helps direct those resources toward customers and problems with stronger commercial evidence.

What should I research first?

Start with the assumption most capable of invalidating the business, such as whether the problem matters, whether a buyer exists, or whether customers can be reached.

What is the difference between primary and secondary research?

Primary research collects new evidence directly through interviews, surveys, observation, and tests. Secondary research uses existing statistics, reports, registers, reviews, and public information.

Should I start with primary or secondary research?

Begin with enough secondary research to understand the market, then use primary research to investigate the exact customer problem and purchasing behavior.

What is qualitative market research?

Qualitative research explores behavior, reasoning, language, processes, and context through methods such as interviews and observation.

What is quantitative market research?

Quantitative research measures frequencies, amounts, preferences, and differences using surveys, transaction data, and statistics.

How many customer interviews should I conduct?

There is no universal number. Ten to fifteen relevant interviews can reveal early patterns, but interviews should continue only while they produce new decision-relevant evidence.

Who should I interview?

Interview people who match the customer definition, recently experienced the problem, influence the purchase, or recently selected or rejected a solution.

What should I ask in a customer interview?

Ask about the last occurrence, current process, consequences, workarounds, spending, purchasing authority, trigger events, and how success is measured.

Should I ask whether customers like my idea?

Do not rely on that question. Ask about actual previous behavior and then present a real offer at a realistic price.

How many survey responses do I need?

The number depends on the population, sampling method, required precision, and intended analysis. A representative sample cannot be created solely by collecting a large number of convenience responses.

Is 100 survey responses enough?

It may provide directional information for a narrowly defined group. It is usually insufficient for precise estimates across several customer segments.

What is a representative sample?

A representative sample reflects the relevant characteristics of the intended population through an appropriate selection and weighting method.

Can I survey my social-media followers?

Yes, but the findings describe the responding followers rather than the entire market unless the audience and sampling method justify a broader conclusion.

How do I research competitors?

Identify direct providers, indirect alternatives, internal workarounds, and inaction. Compare customer, offer, price, delivery, proof, distribution, and limitations.

Is having many competitors bad?

Not necessarily. Competition can indicate established demand and budgets. You still need a credible way to differentiate and reach customers.

What if there are no competitors?

Investigate indirect alternatives and inaction. No visible competition may indicate a new market, weak demand, internal solutions, or regulatory barriers.

How do I determine market size?

Estimate the number of suitable customers, annual spending or revenue per customer, reachable prospects, expected conversion, and owner capacity.

What are TAM, SAM, and SOM?

TAM is the broad theoretical market, SAM is the portion your business can serve, and SOM is the portion it can realistically obtain.

Should I use industry market-size reports?

They can provide context, but their definitions may be too broad. Build a bottom-up estimate using customer counts and realistic business economics.

How do I research customer pricing?

Review public prices, quotations, invoices, alternatives, contracts, and customer purchasing history. Then test a defined offer at a real price.

Can keyword research replace market research?

No. Keyword data measure search behavior. They do not measure all customers, purchases, referrals, or offline demand.

Does high search volume mean a profitable market?

No. Profitability also depends on intent, price, competition, customer acquisition, conversion, delivery cost, and retention.

How do I research a local market?

Use local demographic and business statistics, company registers, maps, local search, reviews, customer interviews, footfall where relevant, and competitor observation.

How do I research a B2B market?

Count purchasing organizations, identify users and approvers, investigate procurement and budgets, analyze company registers, and interview people involved in recent purchases.

How do I research a market with little public data?

Use directories, customer interviews, supplier conversations, manual counting, observed transactions, small sales tests, and transparent assumptions.

Can AI conduct market research?

AI can assist with searching, organizing, transcription, comparison, and analysis. Original sources, customer evidence, and final interpretations still require human verification.

Can I upload customer interviews to an AI tool?

Only after considering consent, confidentiality, personal-data rules, tool retention, security, and contractual obligations. Anonymization may reduce but not always eliminate risk.

How long should market research take?

An initial service opportunity may be researched within two to four weeks. Regulated, technical, manufacturing, or capital-intensive opportunities may require substantially more work.

When is market research finished?

Initial research is finished when you have enough evidence to choose the next responsible action, usually a paid test, offer revision, or decision to stop.

What comes after market research?

Convert the findings into a clear offer and test whether suitable customers will purchase it under realistic conditions.

Key Takeaways

• Market research should support a specific business decision.
• Begin by writing the assumptions that could invalidate the idea.
• Define the exact customer, buyer, geography, and unit being measured.
• Combine secondary research with direct customer evidence.
• Use qualitative research to understand behavior and quantitative research to estimate scale.
• Official statistics provide context but may not match your exact niche.
• Record both the publication date and the reference period.
• Check definitions before comparing figures from different sources.
• Competitors include direct providers, indirect solutions, internal work, and inaction.
• Customer interviews should examine recent behavior rather than hypothetical interest.
• Observation can reveal steps and problems customers omit from their descriptions.
• Surveys are most useful after the important questions are understood.
• A large convenience sample is not automatically representative.
• Pretest questionnaires before distributing them widely.
• Search data reveal language and relative interest rather than total customer demand.
• Google Trends reports normalized relative interest, not absolute search counts.
• Pricing research should compare complete costs and lead to a real priced offer.
• Estimate markets from the bottom up using customer count, price, frequency, access, and capacity.
• A market is useful only when suitable customers can be reached economically.
• Evaluate sources for relevance, accuracy, timeliness, clarity, comparability, and coherence.
• Protect personal and confidential data collected during research.
• AI can assist analysis but should not replace sources, verification, or human judgment.
• Separate observations, patterns, interpretations, and decisions.
• Give stronger weight to purchases and observed behavior than to opinions.
• Define proceed, revise, and stop criteria before reviewing the final evidence.
• Market research should end in action rather than indefinite information collection.

Data and Methodology Note

“Market research” covers several methods with different strengths and limitations.

The evidence and guidance cited in this article draw from:

• U.S. Small Business Administration guidance
• U.S. Census Bureau business and demographic tools
• U.S. Bureau of Economic Analysis consumer-spending data
• Eurostat business, household, and quality statistics
• Companies House public-register guidance
• UK Government Analysis Function questionnaire guidance
• European Commission data-protection guidance
• NIST guidance on generative AI risk management
• Google Trends documentation

Official statistics are not automatically suitable for every business question.

A dataset may be accurate for its stated population while remaining too broad, old, or geographically different for a specific market decision.

Company registers count legal entities according to national registration rules. They may not identify every sole trader, active business location, customer segment, or source of revenue.

Search data measure search activity rather than unique customers or completed purchases.

Interviews provide detailed evidence about experience and behavior but are not normally suitable for estimating population percentages.

Survey results depend on sampling, question design, response patterns, weighting, and analysis. A margin-of-error calculation does not account for every potential source of bias.

TAM, SAM, and SOM calculations are planning estimates. Their quality depends on the definitions and assumptions used.

The research frameworks, checklists, schedules, and evidence rankings are practical editorial tools rather than standardized statistical methods or guarantees of business performance.