Marketing strategy
Market Research: How to Do It Step by Step with Free Sources
The low-budget method: government data that costs nothing, a survey with its margin of error and your market size in dollars before you spend one.
Market research is how you find out, with data and before you invest, whether there are enough customers for your idea, what they'd pay and who you're up against. For a small business it's almost free: the government has already published half the data, and you get the other half with a survey.
We read three of the guides ranking at the top of Google in the US. Many are written by companies that sell surveys or research panels. None names a specific government dataset, none sizes the market with numbers and none tells you how many responses you need before you can trust the result.
Here you do the whole thing, as step one of your marketing strategy: free sources by name, a survey with its margin of error and your market in dollars.
On this page
What is market research (and what it isn't)
Market research is the process of gathering and analyzing data about customers, competitors and the business environment to answer one specific question before you decide. Is there enough demand, at what price and in which area or channel?
If you want what is market research in one line: it's being wrong cheaply on a spreadsheet instead of expensively in a storefront.
What it isn't: it isn't just checking out competitors, it isn't defining your audience and it isn't an 80-page report. It isn't a SWOT analysis either: research brings the data, and the SWOT sorts it so you can decide.
How to do market research in 7 steps
To do market research you need a question, outside data, customer data and a final number. It's seven steps and about a week of evenings.
-
Write down the question it has to answer
"Can I get 90 regular clients at $85 a visit in my part of town?" is a question. "Understand the market" isn't.
-
Define who, where and what they buy
Set the area (a radius, a city or online only), the profile (consumers or businesses) and the problem you solve. Without limits you can't measure anything.
-
Pull free secondary data
Population, income and business counts from the Census Bureau; household spending from BLS; demand trends from Google Trends. Write down every number with its source and date.
-
Study competitors from the outside
Google Maps, Yelp, websites and reviews: how many there are, what they charge and what their customers complain about. Repeated complaints are your opening.
-
Interview 8 to 10 potential customers
Fifteen minutes, open questions: how they solve the problem today, what they spend and what annoys them. This gives you the why.
-
Run a short survey
Eight to ten closed questions in Google Forms, always with a price attached. Aim for at least 100 responses from your audience, not your friends.
-
Calculate your market size and decide
Turn the data into TAM, SAM and SOM and compare it with what you need to sell to break even. If it falls short, change the price, area or product, or drop it.
What market research questions should you ask?
The ones that measure facts and price: how often you buy this, what you spent last time and where. Then, what bugs you about where you buy today and which of three prices feels fair. Keep one open question, at the end. Good market research for small businesses fits on one screen.
Primary vs secondary market research: free data sources
Secondary market research uses data someone else already published, like the Census Bureau. Primary market research is data you collect yourself, with surveys and interviews. Always start with secondary: it's free and fast.
| Source | Type | What you get |
|---|---|---|
| U.S. Census Bureau, Census Business Builder | Secondary | Population, income and business data for the location you pick |
| Census Bureau, County Business Patterns | Secondary | Number of establishments by industry and county |
| BLS, Consumer Expenditure Surveys | Secondary | What households spend, by category |
| Google Trends | Secondary | Whether demand is growing, shrinking or seasonal |
| Google Maps, Yelp and reviews | Secondary | Competitors, prices, ratings and complaints |
| Interviews and your own survey | Primary | The why, current spending and reaction to price |
The SBA's market research guide lists the free federal sources it recommends, from Census Business Builder to the Bureau of Labor Statistics and the Bureau of Economic Analysis. It's the best starting list you'll find, and it's free.
Which free government data should you start with?
Start with three that most guides skip. Census Business Builder gives you people and businesses in your area, County Business Patterns counts your competitors by industry code, and the Consumer Expenditure Surveys tell you what a household spends in your category.
Types of market research and methods
The types of market research cross on three axes: where the data comes from, whether it measures or explains, and what it's for.
| Axis | Type | Example |
|---|---|---|
| Source | Secondary (published data) | Household income in your county |
| Source | Primary (new data) | Survey of 150 neighbors |
| Data | Qualitative (the why) | 10 interviews |
| Data | Quantitative (how many, how much) | Survey with 100+ responses |
| Goal | Exploratory | What local groomers are missing |
| Goal | Descriptive | Average monthly spend per customer |
| Goal | Causal | Do more people book at $79 or at $89? |
The practical rule for market research methods is order. Interviews first, because they tell you what to ask; the survey second, because it tells you how many people think that way.
How many market research survey responses do you need?
With 100 responses your maximum margin of error is about 10 points; with 385, about 5, at 95% confidence. We ran the math:
| Responses | Maximum margin of error | Good enough for |
|---|---|---|
| 50 | ±13.9 points | Checking a problem exists |
| 100 | ±9.8 points | Small, reversible decisions |
| 150 | ±8.0 points | Price and product for a local business |
| 385 | ±5.0 points | A lease, a hire, a big launch |
| 1,000 | ±3.1 points | National or brand studies |
If your audience is small, you need fewer. With 2,000 possible business clients in your state, 323 responses give you the same ±5. But the margin only counts if your audience answered: 100 responses from friends aren't a sample.
Focus groups for market research: how to run one on a budget
A focus group is a guided conversation with 6 to 10 people from your audience, run to learn what they think of your idea and why. Quirk's, the market research trade magazine, calls 90 to 120 minutes the optimum session length. You get your customers' own words and the objections you never thought of.
It doesn't measure anything. Eight people can't tell you how many customers feel the same, only which arguments exist, and the loudest voice sways the rest. Run it before the survey, never instead of it.
A do-it-yourself version:
- Recruit extra people who fit your customer, using two screening questions. Quirk's advises 8 recruits for a group of 6, to cover no-shows.
- Write 10 to 12 open questions, easy ones first.
- Have someone else moderate. If it's your idea, nobody will tell you it's bad.
- Record with permission and note what repeats, not who said it.
- Run at least two groups. Drive Research, a firm that sells them, says one group with one set of people isn't enough.
Here's an invented example. Maya recruits 10 Austin dog owners and expects 8. At $75 each, the low end of what Drive Research calls standard, that's $600. Add $50 of snacks and a free library room: about $650.
| Focus group | In-depth interview | |
|---|---|---|
| Best for | Reacting to a name, a price or a package | Private or complex buying decisions |
| Risk | The loudest voice wins | More hours per person |
| Your time | One evening for 8 voices | About an hour per voice |
Market research companies: when to hire one and how to pick
Hire a research company when a wrong call costs far more than the study. Our rule: if the decision is ten times pricier than the study, pay for it. Otherwise, keep going yourself.
It makes sense when:
- you need a big or national sample;
- your audience is hard to reach, like hospital buyers or CFOs;
- a lender or investor wants an independent view;
- you're too in love with the idea to judge it.
Treat prices as ceilings, because they come from firms that sell the work. Drive Research puts a standard project at $10,000 to $30,000, an online survey of 400 responses at $5,000 to $15,000 or more, and one focus group at $7,000 to $20,000 or more. Older guides quote $4,000 to $6,000 per group, citing Vernon Research Group.
A cheaper first stop is your local Small Business Development Center. America's SBDC says its nearly 1,000 centers offer no-cost business consulting, so ask whether yours helps with market research.
Before you sign a quote, ask in writing:
- The sample size and margin of error.
- How it recruits and screens respondents.
- Whether incentives, facility and moderator are in the price.
- The raw data, not just the slide deck.
- The deadline and how many rounds of changes are included.
How to calculate market size with TAM SAM SOM
Market size comes down to three yearly dollar figures: TAM (all the spending), SAM (what you can serve) and SOM (what you can realistically win).
- TAM: total annual spending on what you sell.
- SAM: the part you can reach with your area, channel and product.
- SOM: what you can win in one to three years, given your capacity and competition.
Top-down starts from a big total and trims it. Bottom-up multiplies customers by yearly spend, and it's the one that works for a small business because it uses your own data. SOM depends on how many interested people actually buy, so your conversion rate matters here: make it up and your SOM is made up too.
Run it in our TAM SAM SOM calculator, which does both methods and writes your pitch sentence.
Market research example: a small business, with numbers
Most market research examples online are big brands. This one is Maya, who wants to launch a mobile dog-grooming van in Austin. It's a made-up case for this guide, with sample figures.
| Step | What she found |
|---|---|
| Question | Can she get 90 clients at $85 a month? She needs $91,800 a year |
| Census data (4 ZIP codes) | 60,000 households, with income above the city median |
| Competitors on Google Maps and Yelp | 6 groomers at $75 to $120, 2 of them mobile; complaints about 3-week waits |
| 10 interviews | Dog owners want weekend slots and no drive to the salon |
| Survey with 150 responses | 8% would book a monthly mobile visit at $80 to $90 |
The math: 8% of 60,000 is 4,800 households, and at $1,020 a year that's a SAM of $4,896,000. But with 150 responses that 8% could be anywhere from 3.7% to 12.3%. Maya plans with the low end: 2,220 households, so her 90 clients would be about 1 in 25.
That's realistic if she owns the gap competitors leave: short waits and weekend slots. Before buying the van, she presells monthly memberships. If 30 people pay, she goes. A "yes, I'd pay" isn't money, and a preorder is the test that doesn't lie.
Act as a market research analyst. My business: [what you sell, where and at what price]. Here's my data with sources: [Census figures, competitors, survey results]. 1) Calculate TAM, SAM and SOM bottom-up and show every multiplication. 2) Separate facts, assumptions and gaps. 3) Tell me which assumption, if wrong, kills the business. Don't invent numbers: if one is missing, ask me for it.
Once your research is done, its conclusions belong at the top of the creative brief you hand your agency or team.
Frequently asked questions about market research
What are the 5 P's of market research?
There's no official list. Most answers borrow the 5 P's of the marketing mix (product, price, place, promotion and people) and treat them as the five areas your research should cover. In practice: what customers want from the product, what they pay today, where they buy, how they find options and who makes the call. If a source gives you a different set, judge it by one test: does it help you ask better questions?
How can I get paid for market research?
By joining paid panels, focus groups or product tests run by research firms, which pay cash or gift cards for specific profiles. Focus groups and in-depth interviews usually pay more than online surveys. Walk away from anyone who asks you to pay to join or wants bank details upfront: legitimate studies pay you, never the other way around. It's a side gig, not the same thing as researching your own market.
What's the difference between market research and marketing research?
Market research studies a market: customers, demand, competitors and size. Marketing research is broader and also studies your marketing: which ads work, which message converts and how people see your brand. In everyday use most people treat them as synonyms, and so do search engines. If you're sizing an opportunity before you launch, you're doing market research.
What are the limitations of market research?
Market research tells you what people say, which isn't always what they'll do. Samples can be too small or skewed, answers go stale as the market moves, and leading questions nudge people toward the answer you wanted. Small niches often have no public data at all. Treat findings as evidence, not a guarantee: check key numbers against a second source, report the margin of error and, when the stakes are high, test with a paid preorder before you commit.
Can ChatGPT do market research?
It can do part of it: draft your survey, summarize interview notes, organize competitor data and run the TAM SAM SOM math on numbers you give it. It can't talk to your customers or know today's local prices, and it may invent precise-sounding figures for small niches. Use it to plan and analyze, and use real customers and government data to validate. Ask it to label every number as fact, assumption or unknown.
How much does market research cost?
Doing it yourself with free sources costs your time plus, at most, a small ad budget to get survey responses. Hiring a firm is a different league: Salesforce, citing Vernon Research Group, puts a single focus group at $4,000 to $6,000. The price depends on sample size, method, how hard your audience is to reach and how many markets you cover. Start with the free route and pay only for what you can't get yourself.
How do you do market research for a new product?
Start with the problem, not the product. Interview 8 to 10 people who have that problem and ask how they solve it today and what it costs them. Then check demand in Google Trends, read competing products' reviews for repeated complaints and test a price with a short survey or a landing page with a preorder button. A paid preorder is stronger evidence than any survey answer.