Small-business market research: a practical 7-day plan

Method in 7 steps

1. 1

   Write one decision the research must support. Examples: choose between two customer segments, understand why estimates are declined, test demand for a new service, select a service area, or decide which benefit should lead the next campaign.

2. 2

   List what is known, what is assumed, and what would change the decision. Keep facts, observations, customer statements, and guesses in separate columns so confidence is visible.

3. 3

   Use secondary research first: public demographic and industry data, local search results, maps, competitor pages, reviews, published prices, directories, associations, and existing business records.

4. 4

   Choose a small primary-research method that answers the remaining question: customer interviews, short surveys, observation, enquiry analysis, lost-estimate review, or a limited message test.

5. 5

   Record evidence consistently. Capture the source, date, geography, segment, question, observation, limitation, and whether the result supports, weakens, or leaves the hypothesis unresolved.

6. 6

   Summarize patterns without pretending a small sample represents the whole market. Write what appeared repeatedly, what contradicted the initial idea, what is still unknown, and which risk deserves another check.

7. 7

   Convert only the validated portion into action: audience, customer problem, proof, offer, channel, message, call to action, and a measurable next experiment. Keep uncertain claims out of public campaigns.

Examples to adapt

• Local roofer: compare public housing and business data with ten recent enquiries, five lost estimates, local competitor positioning, and customer interviews before deciding whether to lead with repair response, inspection, maintenance, or replacement.
• Hair salon: review booking patterns, repeat intervals, cancelled appointments, local reviews, competitor menus, and six customer conversations before promoting color services, quieter weekday slots, or maintenance reminders.
• Restaurant: observe daypart traffic, menu questions, delivery alternatives, local office patterns, reviews, and a short customer survey before creating a lunch, takeout, event, or loyalty campaign.
• Accounting firm: analyze enquiry sources, seasonal questions, service mix, business counts by industry, competitor promises, and client interviews before selecting a niche message or educational campaign.
• AdSpark workflow: enter the evidence-backed business moment, audience, proof, limitations, and next step; generate channel-specific drafts; then verify every statement before scheduling or publishing.

Start with one business decision, not a pile of data

Useful research reduces uncertainty around a real decision. `Who are our customers?` is too broad to complete. `Should we promote emergency repair or preventive inspection to homeowners in our current service area next month?` can be tested. The decision determines which population, alternatives, behavior, geography, time period, and evidence matter.

Write a hypothesis before searching, but do not write the conclusion. A good hypothesis names the segment, problem, expected behavior, and evidence that would change your mind. This prevents the owner from collecting only facts that support the preferred idea.

Combine secondary research with direct customer evidence

Secondary research uses information that already exists. It is efficient for market size, demographics, business counts, industry classification, employment, prices, geography, competitors, trends, and observable customer language. Primary research collects evidence for the business's specific question through interviews, surveys, observation, service records, experiments, or other direct contact.

The U.S. Small Business Administration recommends combining market research and competitive analysis. It separates broad, quantifiable existing sources from direct methods such as surveys, questionnaires, focus groups, and in-depth interviews. For a small business, the practical sequence is usually public data first, focused direct research second, and a limited test third.

Use free U.S. data without turning estimates into facts

Public sources can establish the context around a local or small-business market. The SBA points owners to the North American Industry Classification System, Census Business Builder, data.census.gov, the Bureau of Labor Statistics, the Bureau of Economic Analysis, the Federal Reserve, and other federal sources. Use the source that matches the question and record the dataset date, geography, unit, industry code, and limitations.

Census Business Builder combines demographic, economic, and business data for selected industries and locations. It can help compare areas and understand the environment around a business, but it does not identify which individual person will buy or guarantee demand. Public counts should inform an estimate range and a test plan, not become a fabricated total addressable market slide.

A seven-day market research plan for a busy owner

The goal is not to complete a permanent market study in one week. It is to replace the weakest assumptions behind one decision with enough evidence for a responsible, limited next step. Schedule short blocks and keep an evidence log so another person can understand where each conclusion came from.

Do not wait for perfect data. Define what uncertainty is acceptable for a reversible marketing test and what uncertainty is too dangerous for pricing, hiring, inventory, financing, health, legal, safety, or other high-impact decisions. Those decisions may require qualified professional research or advice.

Ask customer questions that reveal behavior, not compliments

Avoid asking whether someone likes an idea or would hypothetically buy it. Ask about a recent situation: what happened, what triggered the search, what they tried, which alternatives they considered, what nearly stopped the purchase, what evidence helped, and what happened after the decision. Specific past behavior is usually more useful than praise for a concept.

Use the same core questions across interviews, then follow relevant details without leading the person. Do not recruit only your happiest customers. Include different outcomes where appropriate: new customers, repeat customers, declined estimates, abandoned enquiries, former customers, and people who chose an alternative.

• Tell me about the last time you tried to solve this problem. What happened first?
• What made the issue important enough to act on that day or week?
• What did you try before contacting or choosing a business like ours?
• Which other options did you consider, including doing nothing or doing it yourself?
• What information or proof did you look for before deciding?
• What was confusing, risky, slow, expensive, or frustrating about the process?
• Where did you search, who did you ask, and which sources did you trust?
• What nearly stopped you, and what finally made the next step feel reasonable?
• What outcome mattered after the purchase, and how did you judge whether it worked?
• What should we have asked but did not?

Compare the whole customer alternative—not just competitor websites

A competitor matrix should include direct competitors, specialists, broad providers, DIY options, software, marketplaces where relevant, delaying the decision, and doing nothing. Record only what is observable: the audience named, problem described, offer, public price or pricing model, proof, process, channel, response path, reviews, locations, constraints, and update date.

Do not infer revenue, customer counts, conversion, service quality, profitability, or market share from a polished site. Treat review themes as qualitative signals, not verified representative statistics. The objective is to find unanswered customer questions and credible differences—not to copy claims or produce a fictional SWOT analysis.

Collect less personal data and preserve the evidence trail

Explain why information is being collected, who will use it, whether participation is voluntary, and how notes will be stored or shared. Avoid collecting sensitive or identifying information unless it is genuinely necessary and there is an appropriate lawful, secure process. Requirements vary by jurisdiction, audience, method, and data type; confirm the rules that apply to the business.

The Federal Trade Commission's business guidance recommends knowing what personal information the business holds, keeping only what is needed, protecting retained data, disposing of it properly, and planning for incidents. A small research project does not justify creating a permanent file of unnecessary personal details. Anonymize working notes when identities are not needed and define a deletion date.

• Collect the minimum information needed to answer the decision question.
• Separate contact details from research notes when possible and restrict access.
• Do not publish interview words, names, photos, recordings, or identifiable details without appropriate permission.
• Record consent, source, collection date, retention decision, and any limits on reuse.
• Delete raw information when the business purpose and required retention period end.

Turn research into a campaign brief without overstating the findings

The final research output should be short enough to use. Name the decision, segment, problem, trigger, alternatives, exact customer language, strongest proof, unresolved objection, source trail, confidence, and next test. Then define the campaign moment, channel role, destination, call to action, measurement, and stop rule.

AdSpark can use the evidence-backed brief to prepare ready-to-review versions for email, social, Google Business, and other supported channels. It can preserve business context and reduce rewriting across channels. It cannot conduct representative sampling, verify market size, access private competitor data, infer customer consent, or certify that a conclusion is statistically reliable. The owner must review facts, permissions, offer details, audience, and publication.

Common questions before publishing