Primary Market Research: A Practical Guide for B2B and SaaS Teams
Methods, sample sizes, costs and a seven-step process for B2B and SaaS teams, plus why most B2B marketing decisions still skip it.
Updated November 5, 2026
Primary market research is research you run yourself to answer your own question, by collecting new data directly from the people you want to sell to. Surveys, interviews, message tests and brand tracking are all primary research. Reading an analyst report or a competitor's website is secondary research: someone else's data, collected for someone else's question.
In B2B, primary research is how you find out what your buyers think instead of guessing. Most teams guess. I've watched 1,800+ B2B companies run research on Wynter, and the gap between teams that ask and teams that assume shows up everywhere, from homepage copy to launch plans. Here's what primary market research is, which methods work in B2B, what it costs, and how to run it without wasting the budget.
| Primary research | Secondary research | |
|---|---|---|
| What it is | Data you collect yourself, for your question | Data someone else already collected |
| Examples | ICP surveys, interviews, message tests, brand tracking | Analyst reports, industry studies, competitor sites, desk research |
| Answers | What your buyers think, in their words, right now | What the market looks like on average |
| Speed | Hours to weeks | Minutes to days |
| Cost | From a few hundred dollars to six figures | Free to a few thousand dollars per report |
| Biggest risk | Asking the wrong people | Data that is old, generic or about someone else's buyers |
You need both. Secondary research tells you the shape of the market and what's already known. Primary research tells you what your specific buyers think about your specific problem, product and message. If two competitors read the same Gartner report, they get the same insight. If one of them surveys 100 buyers in their ICP, they get something the other one doesn't have. We go deeper on the trade-offs in primary vs. secondary market research.
Pick the method based on the question, not on what you used last time.
| Method | What it answers | Typical sample | B2B watch-out |
|---|---|---|---|
| ICP surveys | How buyers describe their pains, triggers, shortlists and criteria | 30 to 50 for open-ended answers, 100+ for percentages | Unverified panels full of fake "directors" |
| Interviews | Why buyers act the way they do, in depth | 9 to 17 per segment | Recruiting senior people takes time |
| Message tests | Whether your copy is clear, relevant and believable to your ICP | 15 to 30 | Testing before you know the buyer |
| Preference and concept tests | Which option buyers pick, and why | 30 to 50 | Close results are ties, not wins |
| Brand tracking | Awareness and consideration over time | 100 to 200 per wave | Small brands move less than the margin of error |
| Win-loss interviews | Why you won or lost real deals | Every meaningful deal | Reps running them get polite answers |
| Focus groups | Group reactions to early ideas | 6 to 8 per group | Hard to schedule senior B2B people together, and one loud voice skews the room |
Two rules of thumb. Go qualitative first to find the right questions, then quantitative to size the answers (more in qualitative vs. quantitative research). And for anything you'll report as a percentage, get at least 100 responses.
The questions B2B marketing and product teams bring to us most often:
One sentence: "Depending on what we learn, we will ___." Rewrite the homepage. Pick a segment. Kill a feature. If the answer can't change a decision, don't run the study. Research without a decision attached becomes a deck nobody opens.
Role, seniority, company size, industry. "Marketing leaders" is not an audience. "VPs of Marketing at B2B SaaS companies with 200 to 2,000 employees" is. Then decide which of four groups you need: current customers, lost deals, in-market non-customers or out-of-market category buyers. Most teams only research the first. Our B2B customer research playbook explains what each group can and can't tell you.
Use the table above. If you don't yet know which questions matter, start with interviews or open-ended survey questions. If you need to choose between options, use a preference test. If you need a trend, use tracking.
This is where B2B primary research usually breaks. Your own email list is people who already like you. Generic online panels are full of people pretending to hold senior titles, because senior titles pay the highest incentives. Ask any vendor how they verify job title and employer. "They self-report it" is not an answer. Our guide to B2B research panels has the full vetting checklist.
"Walk me through the last time you bought one of these" beats "What matters most to you?" People are poor narrators of their own decisions and worse predictors of their future ones. Our list of 40 qualitative research questions is built on that rule, and our 60 B2B market research survey questions are grouped by the decision they inform.
Code open-text answers into themes, look for differences between segments, and write down what the finding changes, who owns it and when to revisit it. Here's a decision-record template you can copy. Then repeat on a schedule. Buyer knowledge expires.
| Option | Rough cost | Speed | Good for |
|---|---|---|---|
| Your own list (DIY) | Your time plus incentives | Weeks | Customer research. It is a fan club, not the market. |
| Per-complete panel providers | Roughly $50 to $90 per survey response for professional audiences, more for senior roles | Days to weeks | Teams that design and analyze research themselves |
| Self-serve B2B research platforms | Annual credit plans or pay-as-you-go per test | 12 to 48 hours on Wynter | Repeated studies with the same ICP |
| Expert networks | Priced per call, and senior experts are expensive | Days | A handful of deep conversations with rare experts |
| Full-service agencies | Often $50,000 to $250,000+ per study | Weeks to months | Large, multi-country, high-stakes programs |
The agency numbers aren't hypothetical. When Maxio wanted a brand awareness study of its ICP, agencies quoted $100,000 to $240,000. They ran it on Wynter for a little over $10,000 and re-ran it a year later to measure the lift.
Don't optimize for the cheapest response. A cheap answer from someone who isn't who they say they are is the most expensive data you'll buy, because you'll act on it. See Wynter pricing for how credits work.
SaaS and tech market research is the same discipline with a few twists. The buyers are harder to reach, the market moves faster and most of the vendor shortlist now forms before anyone fills in a demo form. In our 2026 study of how B2B SaaS CMOs buy software, 84% said they use AI tools like ChatGPT, Claude and Perplexity for vendor discovery, up from 24% a year earlier.
| What's different | Why it matters | What to do |
|---|---|---|
| Buyers | Technical, skeptical, hard to reach. Many are engineers, IT and security people who ignore cold outreach. | Survey by exact role and seniority on a verified panel, not a generic "business professionals" pool. |
| Speed of change | Categories, competitors and pricing shift within a year, now faster with AI. | Treat research older than 12 months as stale. Re-run core surveys yearly. |
| How the shortlist forms | Buyers research vendors in AI assistants, peer communities and review sites before they talk to you. | Ask where they looked and which vendors came up, not only what they think of you. |
| Product data | Usage analytics show what customers do, not why non-customers never signed up. | Pair product data with research on people who never became customers. |
| Sameness | Most SaaS homepages make the same claims, so buyers struggle to tell vendors apart. | Test your messaging with buyers before you ship it, including with the logo hidden. |
The questions SaaS teams bring to primary research most often: which segment to go after next, why trials don't convert, how buyers compare you with two or three named competitors, and whether a new positioning is clear to people who've never heard of you. Each one is answered by people outside your customer base, which is why tech companies can't rely on product analytics and customer calls alone.
More in our roundup of actionable insights from B2B marketing leaders.
Research where you collect new data directly from your target audience, for example through surveys, interviews, message tests or brand tracking, to answer a question specific to your business.
An ICP survey on buyer pains, customer interviews, win-loss interviews, a message test of your homepage, a preference test between two positioning options, focus groups and brand tracking surveys.
From a few hundred dollars for a small self-serve test to six figures for a full-service agency study. Cost is driven mainly by audience seniority, sample size and how much of the work you do yourself.
Self-serve B2B tests can return results in 12 to 48 hours. Interview programs usually take two to six weeks. Agency studies often take one to three months.
Neither is better. Secondary research is faster and cheaper for general market facts. Primary research is the only way to learn what your specific buyers think about your specific product and message.
Market research for software companies: surveys, interviews and tests with the buyers and users of a software category, used to choose segments, shape positioning and messaging, plan the roadmap and track brand awareness. It differs from other B2B research mainly in how fast the market changes and how technical and hard to reach the buyers are.
Your competitors read the same reports you do. What they usually don't have is fresh answers from the buyers you're both selling to. If you want verified B2B buyers from your ICP answering your questions in 12 to 48 hours, start with a Wynter survey, or follow the Wynter Research Playbook for what to run first.
.jpg)
.png)