You can do solid market research for a small business without hiring a firm by combining free public data, a close study of competitors, a handful of real customer interviews, a short survey and a small paid test of demand. The goal isn't a thick report. It's a reasonable estimate of how many customers exist, what they want and what they'll pay.
This guide is for owners planning a new business, product, location or service line who want evidence before they commit money. You'll learn which research methods answer which questions, how to size a market from the bottom up with real arithmetic, an interview script that avoids misleading answers, and a two-week research plan you can follow.
What market research should answer
Before gathering anything, write down the decisions the research needs to support. Most small-business research boils down to five questions:
- Who exactly is the customer? Not "everyone," but a specific group you can describe and reach.
- How many of them are there within your reach?
- What problem are they trying to solve, and how do they solve it today?
- What will they pay, and how often?
- Who else serves them, and where are the gaps?
If a piece of research doesn't help answer one of these, skip it. Our guide on finding your target market goes deeper on the first question.
Research methods that cost little or nothing
| Method | Best for answering | Cost | Watch out for |
|---|---|---|---|
| Public data from government and local economic development offices | How many potential customers exist in your area | Free | Data may be a few years old. Use it for scale, not precision. |
| Trade and industry associations | Industry norms, trends, typical costs | Free to modest membership fees | Figures may reflect larger businesses |
| Competitor research: websites, prices, reviews | What's offered, at what price, and what customers complain about | Free | Posted prices may not match what's actually charged |
| Customer interviews | Problems, current solutions, buying process | Your time, maybe a small thank-you gift | Small samples. Look for patterns, not percentages. |
| Short surveys | How common a need or preference is | Free to low | People overstate what they'll buy |
| Observation: counting foot traffic, visiting competitors | Real behavior at a location | Your time | Pick representative days and times |
| Demand test: a pre-order page or small ad campaign | Whether people act, not just say | A few hundred dollars | Make sure the offer is honest about timing |
The single richest free source is often competitor reviews. Read fifty reviews across three or four competitors and list every complaint and every compliment. Recurring complaints are gaps you might fill. Our guide on doing a competitive analysis gives you a structured way to compare them.
How to run customer interviews that tell you the truth
People are polite. Ask "Would you buy this?" and many will say yes to be nice. The fix is to ask about past behavior and real spending, not hypothetical intentions.
An interview script you can adapt
- "Tell me about the last time you [dealt with the problem]. What happened?"
- "How do you handle it today? What do you use or who do you pay?"
- "What do you spend on that, roughly, in a month?"
- "What's the most annoying part of how you do it now?"
- "Have you looked for a better option? What did you find, and why didn't you switch?"
- "If something solved that, what would it need to do for you to switch?"
- "Who else should I talk to about this?"
Notice that the script never describes your idea until the very end, if at all. You want to hear their world in their words. Aim for ten to fifteen interviews with people who fit your target customer. When you start hearing the same answers repeatedly, you've learned most of what interviews can teach.
Worked example: sizing a meal-prep delivery market
Suppose a hypothetical entrepreneur wants to launch a weekly meal-prep delivery service in a mid-sized town. They need to know whether there are enough potential customers to support it. Here's a bottom-up estimate, which starts from real, countable customers rather than a big national figure.
- Households in the delivery area: public data shows about 40,000.
- Households that match the target, busy two-income families with children, based on the same public data: suppose about 9,000.
- Survey result: of 150 target-fit respondents, 27, or 18%, said they were "very likely" to try a service like this.
- Adjust for overstatement: because stated intent runs high, the owner halves it to 9%. That gives 9,000 × 0.09 = 810 realistic prospects.
- Share they can realistically win in two years: with limited marketing and competition from grocery delivery, suppose one in four, so 810 ÷ 4 ≈ 202, rounded to 200 customers.
- Revenue per customer: interviews suggest two $85 orders a month, so $170 a month.
- Annual revenue potential: 200 × $170 × 12 = $408,000.
Then compare with what the business needs. If the owner's cost model shows they need 150 active customers to break even, that's $306,000 a year, and the estimate of 200 leaves some margin for error. If the model needed 400 customers, the idea would need a bigger area, a higher price or a different customer group.
Every step here is an assumption, so test the fragile ones. The halving of stated intent and the one-in-four share are guesses. That's where a demand test comes in.
Testing demand with real behavior
The owner builds a simple page describing the service, the price and a planned launch date, with a "reserve your first week" sign-up. A $300 local ad campaign sends 1,500 visitors to it, and 45 people sign up.
- Sign-up rate: 45 ÷ 1,500 = 3%
- Cost per sign-up: $300 ÷ 45 = about $6.67
Forty-five real sign-ups from a small test is much stronger evidence than 18% of a survey saying "very likely." Be honest on the page about the launch timing, and follow up with everyone who signs up. If you're testing an idea before leaving a job, our guide on validating a business idea before you quit covers more tests like this.
A two-week market research plan
| Days | Focus | Output |
|---|---|---|
| 1–2 | Write your five questions and your assumptions | One page listing what you need to learn |
| 3–4 | Public data and industry sources | Rough count of potential customers in your area |
| 5–6 | Competitor research and review reading | A table of competitors, prices, strengths and common complaints |
| 7–11 | Ten to fifteen customer interviews | Notes with recurring problems, current spending and switching triggers |
| 8–12 | Short survey to a wider group | How common the key needs and preferences are |
| 10–14 | Demand test | Real sign-ups, inquiries or pre-orders |
| 14 | Pull it together | Bottom-up market estimate and a go, adjust or stop decision |
Keep surveys short, five to eight questions, and include at least one question about current spending or behavior. Ask for honest answers, and don't reveal which answer you're hoping for.
Common market research mistakes
- Starting with a huge top-down number. "The national market is billions, so if we get 1%..." tells you nothing about whether you can win customers. Build up from customers you can count and reach.
- Asking friends and family. They want to support you, so their answers skew positive. Talk to strangers who fit your target.
- Leading questions. "Wouldn't it be great if…" invites a yes. Ask about past behavior and let people describe problems in their own words.
- Treating stated intent as sales. What people say they'll buy runs higher than what they actually buy. Discount it, and test with real actions.
- Ignoring the "do nothing" option. Your biggest competitor is often the customer's current habit. Find out why they haven't changed already.
- Researching forever. Research reduces risk but never removes it. Set a deadline and a decision rule in advance.
Where AI can help
AI tools are useful for processing what you collect. Paste in anonymized interview notes and ask an assistant to identify recurring themes and notable quotes. Export competitor reviews you've gathered and have it categorize complaints and praise. It can also check your survey questions for leading wording and help build the bottom-up sizing model in a spreadsheet so you can change assumptions easily.
Don't let it replace the research itself. AI tools can produce plausible-sounding market figures that aren't grounded in real data, so never use a number from an AI tool unless you've traced it to a source you trust. The interviews and demand tests are where the real insight comes from.
Frequently asked questions
How much should a small business spend on market research?
Often very little in cash. Most useful methods cost time rather than money, and a demand test might cost a few hundred dollars. Scale the effort to the size of the decision: a new location deserves more research than a new menu item.
How many people do I need to interview?
Ten to fifteen well-chosen interviews often reveal the main patterns. If new interviews keep surfacing different problems, keep going, because your target group may be too broad.
Is a survey enough on its own?
Usually not. Surveys tell you how common a stated preference is but not the reasons behind it, and they overstate buying intent. Combine them with interviews and a test that measures real behavior.
When should I hire a research firm?
Consider one for large, high-stakes decisions, specialized markets you can't reach yourself or when you need statistically reliable results. For most small-business decisions, the methods here are enough.
Key takeaways
- Start by writing the specific decisions your research needs to support.
- Combine public data, competitor reviews, interviews, a short survey and a demand test.
- Ask about past behavior and current spending, not hypothetical intentions.
- Size your market bottom-up from customers you can count and reach, and compare it with your break-even need.
- Discount stated intent, and trust real sign-ups and pre-orders over survey answers.
Comments
Post a Comment