By the Wavefield Research team · Published Aug 31, 2026
Market research costs anywhere from nothing to six figures for the same question, so the useful answer is by tier: free using your own list, $99 to $500 per study on a self-serve platform, $1,000 to $5,000 once you buy panel sample, and $15,000 or more for a full-service agency.
Which tier you pick decides the price far more than anything you negotiate inside it, and that is the part most guides to market research pricing leave out. Most published cost guides are written by agencies and start their ranges at about $5,000, because that is where their own work begins. This one covers the tiers below that line too, decomposes what is actually inside a five-figure quote, and says plainly when paying an agency is the right call. We sell a platform in the second tier, so read the agency section as the honest limit of our own pitch.
The same research question can be answered at any of these levels. What changes is how much of the work you do, how defensible the numbers are, and who is accountable for the conclusion.
Swipe the table sideways to see all columns →
| Tier | Typical cost | What you get | What you give up |
|---|---|---|---|
| Free tools, your own audience | $0 | Google Forms or a free plan, sent to a list you already have. Counts and charts. | No sampling control, no weighting, no significance testing. Fine for internal questions, indefensible for a decision anyone will challenge. |
| Self-serve research platform | $99–$500 per study | A real instrument — screeners, quotas, logic — with weighting and significance-tested reporting. You bring or buy the respondents. | Your own time to review the questionnaire and read the results. No account team, no norms database, no strategic recommendation. |
| Platform plus sourced sample | $1,000–$5,000 per study | The above, plus panel sample bought for the study. This is where most small and mid-size businesses can actually operate. | Still self-directed. Panel cost dominates the total, and it scales with how rare your audience is. |
| Full-service agency | $15,000–$75,000+ per study | Design, fielding, analysis, and a presented deck, run by researchers who own the outcome. Complex methods, specialist audiences, and defensibility for high-stakes decisions. | Budget and calendar time — typically six to twelve weeks, and a quote instead of a price. |
Agency figures are commonly published ranges as of August 2026 and vary widely by audience and market — treat them as orientation, not quotes. The highlighted row is where Wavefield sits: $99 to $299 per project, published on the page, with respondents supplied by you or sourced separately.
Quotes arrive as one number, which makes them impossible to evaluate. Most of the cost of market research is labor, and labor divides into parts you can price separately. A standard five-figure online study splits roughly like this — proportions are approximate and vary by firm, but the shape holds:
The irreducible part. You are buying people's attention, and no software removes it.
Turning the brief into a working instrument — logic, quotas, rotations, testing. Automatable.
Scheduling, panel liaison, status calls, quota monitoring during fieldwork. Partly automatable.
Weighting, crosstabs, significance testing, chart building, the deck. Largely automatable up to the interpretation.
Offices, business development, and the risk of running a services firm. Real costs, but yours to decide whether to buy.
Read the column on the right and the strategy falls out. Roughly half of a typical quote is labor that software now does: programming the instrument, watching quotas fill, computing weights and crosstabs, building tables. The sample is irreducible, and the interpretation — what the numbers mean for your decision — is the part worth a human. A team that buys its own sample and automates the middle can run the same study for a fraction, and should be honest that it is now doing the project management itself.
Totals hide everything. A $15,000 study of 1,000 people and a $15,000 study of 200 specialists are different purchases, and only the per-complete figure says so. Two benchmarks worth carrying into any conversation, both commonly reported ranges rather than fixed prices: panel sample for general-population online research runs a few dollars to low double digits per completed interview, while full-service agency work is often quoted between $30 and $150 per complete all-in.
The distance between those two numbers is the service layer, and naming it changes the negotiation. You are no longer haggling over a total; you are deciding how much design, management, analysis, and accountability you want to buy on top of the respondents. Sometimes the answer is all of it. Often, for a straightforward tracker or a concept test, it is not.
One caution when comparing quotes: confirm what happens to screen-outs. Panels bill for people who enter and fail your screeners as well as those who finish, so a low-incidence study can cost far more than the per-complete headline suggests. That is the incidence math doing its work, and it is the most common reason a quote gets renegotiated mid-field.
The single biggest multiplier. At 50% incidence you screen two people per complete; at 5% you screen twenty, and the panel prices accordingly. A study of general consumers and a study of left-handed CFOs use the same software and differ tenfold in cost. The math is in survey screening questions.
Cost scales almost linearly with completes, while precision improves with the square root — going from n=400 to n=1,600 quadruples sample spend to halve the margin of error. Decide the margin you need first with the sample size calculator, then buy exactly that.
Panels price by length as well as difficulty; past roughly fifteen minutes both the incentive and the dropout rate climb. Cutting a questionnaire from twenty-five minutes to twelve is often the cheapest saving available.
Consumers are cheap, B2B decision-makers are expensive, and named-account executives may not be buyable at any price. Specialist audiences are where agency recruiting genuinely earns its fee.
Online is the cheap default; phone multiplies cost several times over for interviewer labor; in-person more again. Mode also changes who answers, which is a data-quality question before it is a budget one.
Tables and an export cost near nothing once the data exists. A written narrative, a designed deck, and a presented readout are labor, and they are frequently the difference between a $6,000 quote and a $15,000 one.
We sell software, so treat this as the boundary of our own argument. There are studies where the five-figure quote is the cheap option:
Hospital procurement leads, farm equipment buyers, C-suite in a named industry. Recruiting these people is relationship work and specialist panel access, and a firm that already has them saves you months of failing to find them.
Conjoint with a large attribute space, MaxDiff, segmentation that has to survive a statistician's review. Software can field these; designing them badly is expensive in a way that does not show up until the results mislead you.
A $40,000 study protecting a $10 million launch is insurance, not overhead. When being wrong is catastrophic, buy the accountability of a firm whose name is on the recommendation.
Self-serve assumes someone will read the crosstabs and write the story. If no one has that time, an unread $99 study costs more than a presented $20,000 one — the cheap option is only cheap if it gets used.
The mirror image is equally true. A recurring tracker, a concept test between two options, a customer satisfaction read, a district poll — these are well-defined studies with standard instruments, and paying agency rates for them buys mostly project management. That is the work automated questionnaire programming and built-in analysis are for. Where each method sits on cost and speed is mapped in market research methods, the honest map.
It spans four orders of magnitude, so the honest answer is by tier: free with your own audience and free tools; $99 to $500 per study on a self-serve research platform where you supply respondents; $1,000 to $5,000 once you buy panel sample for a typical consumer study; and $15,000 to $75,000 or more for a full-service agency project. Published agency ranges for a standard online survey commonly start around $5,000 and rise from there, which is why the tier you pick matters more than any negotiation within it.
Do the parts that are structured, and buy the parts that are not. Writing a clear brief, running a survey on a self-serve platform, and reading a weighted topline are all within reach of a non-researcher, and that path costs $99 to $500 per study before respondents. Buying panel sample takes a typical study to the low four figures. What to be careful about doing yourself: screener design, quota setup, and question wording, where mistakes are invisible until the numbers mislead you — a platform that programs those from a brief removes most of that risk.
Because most agency work is genuinely custom — incidence, audience, and length move the real cost several-fold — and because quoting per project lets pricing follow the client's budget rather than the study's cost. Both reasons are legitimate; neither helps you plan. The practical response is to ask for the quote broken into sample, programming, analysis, and project management, so you can see which parts you are buying and which you could do yourself.
Judge it per complete rather than in total. Panel sample for a general-population online study commonly runs a few dollars to low double digits per completed interview, rising steeply for low-incidence audiences. Full-service agency work is often quoted at $30 to $150 per complete all-in, and the gap between those figures is the service layer — design, management, analysis, and margin. Neither number is wrong; you are deciding how much of that layer you need.
Yes, if you scope honestly. A market research budget in the low four figures covers a focused study on a general-population audience with a short questionnaire and a few hundred completes, and the cost falls to almost nothing if you already have a customer list to survey. What a small budget cannot buy is a rare audience, a long questionnaire, and a strategic narrative at the same time. Cut the audience difficulty and the length first; cut the sample size last, since that is what your margin of error is made of.
Related: market research methods · market research surveys, type by type · online survey platforms compared
$99 to $299 per project, free to build, no contract. The agent programs the questionnaire; weighting, crosstabs, and exports come with it. You bring the respondents or source panel sample separately.
Start a study