Market research for a small business runs on a short ladder: free existing data to frame the market, a cheap survey of your own contacts, panel respondents when you need strangers, and an agency almost never. This guide walks the ladder with real prices on every rung, plus the startup version and the case for sometimes skipping research entirely.
By the Wavefield Research team · Published Oct 22, 2026
Most guides to small business market research open with a tour of research types. That ordering is backwards, and it is why so much small business research ends as a folder nobody opens. Start by writing down the decision: should I open the second location, raise the price, add the service, target the new segment? A decision tells you exactly what information would change your mind, which tells you what to collect and, just as useful, what to ignore. Research that is not attached to a decision is a hobby. A pleasant one, but your competitors would prefer you kept at it.
Before spending anything, take what is free, because it is better than its price suggests. The SBA's market research guide lists the federal sources, and three of them do most of the work. Census data tells you how many people of what age and income live in your trade area, which sizes a market more honestly than any gut feel. Google Trends shows whether demand for what you sell is growing, shrinking, or seasonal, and how your region compares. And your competitors' reviews are free qualitative research that someone else paid to collect: read a hundred of them and you will know the category's unmet complaints better than most consultants. The free stack answers the framing questions, who is out there, what they already buy, and what they complain about. What it cannot answer is anything about your specific product, your price, or your concept, because no existing dataset contains your idea. One practical tip for the review reading: sort critical reviews first and tally complaints into five or six buckets as you go, because the bucket counts are the finding, and three-star reviews usually hold the most honest detail. Past that line, research starts costing money, and our market research methods guide maps everything on both sides of it.
Here is the part the other guides skip: numbers. Every guide on page one of this topic says research matters and none says what it costs, which for a small business is the entire question.
Census demographics, Google Trends demand curves, competitor reviews read closely, SBA data sources. Answers who is out there and what they already buy.
A real survey, built by an AI agent from your brief, fielded to your customer list or community by email link. Answers what your existing audience wants next.
People you don't know yet, matched to your target market with quotas. 200 completes lands around $800 to $1,000 all in. Answers whether strangers would buy.
The study designed, run, and presented for you. Right when the decision is big enough to justify it, which for most small businesses is rarely.
To make the ladder concrete, here is what a real study looks like for a bakery owner weighing a second location across town. The decision: open it or not, with a lease that costs real money to exit. The free stack first: census data says the new neighborhood has the right density and incomes, and competitor reviews reveal one underperforming incumbent. That framing is worth exactly what it costs, and it cannot say whether those residents would switch. So, rung three: a 12-question survey of 200 panel respondents quota'd to the new area, asking about current bakery habits, what would make them switch, price sensitivity on the three core products, and a direct concept question about the planned location. Cost: $99 study plus 200 completes at $3.50, so $799 total, back in under a week. The crosstab that matters splits intent by distance, because “would visit” from someone eleven minutes away is a different fact than from someone three minutes away. Whatever the answer, $799 against a five-year lease is the cheapest insurance in the building. That is the whole method: decision, framing from free data, one focused study priced in hundreds, and a readout aimed at the specific numbers that would change your mind.
Full tier-by-tier decomposition, including what agencies actually charge for and when they earn it: how much market research costs. What to field once you pick a rung: market research surveys, type by type.
Market research for a startup uses the same ladder with the middle rung sawed off: there is no customer list to survey cheaply, which is exactly why startup founders get quoted the agency rung. The workable substitutes, in the order to try them: problem interviews (ten real conversations with people who have the problem, none of whom are your friends), competitor evidence (their reviews, their pricing pages, their job postings tell you what is working), and then a small concept test on panel respondents, where a few hundred dollars buys reactions from strangers who owe you nothing. The founder-specific trap deserves its own sentence: the people you can reach free are the people who like you, and a concept validated by friends has survived nothing. If the test matters, pay for indifference. For the concept test itself, quotas on category usage matter more than demographics, and the mechanics are covered in our concept testing guide. Startup teams that make research a monthly habit get more from trackers than from one-off studies, but that is a habit to build after revenue, not before.
No research vendor says this, so it falls to us: sometimes the honest answer to “should a small business market research this” is no. The test is reversibility. If trying the thing in the real market costs less than studying it, the market is the better instrument: run the weekend special, change the price for a month, ship the ad variant, and read actual behavior instead of stated intentions. Research earns its keep on decisions that are expensive to undo, a lease, a build, a rebrand, a hire, because there the cost of being wrong dwarfs the cost of asking first. Small business owners mostly err in both directions at once: researching the reversible decisions out of nervousness and winging the irreversible ones out of impatience. Flip that and you are ahead of most of the market without spending a dollar more.
In order: write down the one decision the research must inform, spend a few free evenings on existing data (census demographics, Google Trends, competitor reviews), and only then pay for primary research aimed at your exact question, starting with a survey of your own contacts before buying panel respondents. Most small business research fails at step one, by gathering information no decision was waiting for.
The honest range is $0 to about $1,000 for almost everything a small business needs. Secondary research is free. A surveyed study of your own customers runs $99 on a per-project platform. Reaching 200 matched strangers through a panel lands around $800 to $1,000 total. Agency studies start around $15,000 and exist for decisions with commas in them.
Genuinely yes, for the framing half. Census data sizes your demographics, Google Trends shows demand direction, competitor reviews are free qualitative data, and the SBA lists more sources. The free stack cannot answer questions about your specific product or price, because that takes real respondents, and respondent time is the part that costs money.
Same tools, nastier trap. A startup usually has no customer list, so the cheap middle tier is missing, and the people founders can reach free (friends, other founders) are the worst possible sample because they want you to succeed. Market research for startup companies leans on problem interviews, competitor evidence, and small panel-sampled concept tests rather than surveying the people who already like you.
Fewer than vendors imply, more than a group chat. For a directional read on one audience, 100 to 200 completes gives a margin of error around 7 to 10 points, which is enough to choose between two concepts but not to detect small shifts. Run your numbers in a sample size calculator before fielding, so the budget matches the decision.
When the decision is cheap and reversible. If testing in the real market costs less than studying it (a weekend menu item, a price tweaked for a month, one ad variant), ship the test and read the till. Research earns its cost when the decision is expensive to reverse: a lease, a product line, a rebrand, a second location.
Size the study before you buy it: sample size calculator · Which software category you actually need: market research platforms, the six kinds
Describe your study and the AI agent builds it. Field to your own contacts from $99, or reach matched panel respondents from $3.50 per complete. You don't pay until you field.
Start a project