How Much Does a Market Research Survey Cost?

1. Incidence: how many people qualify

Incidence is the share of people who enter your screener and qualify for the survey. It is usually the single biggest driver of sample cost, because a provider has to invite and screen everyone who doesn't qualify in order to find those who do.

Illustrative example (made-up numbers): to get 300 completes at 50% incidence, roughly 600 people need to be screened. At 10% incidence, roughly 3,000 need to be screened. Same number of completes, five times the screening effort.

2. Sample size: how many completes you need

More completes cost more, but precision doesn't improve in a straight line. For a simple random sample, the margin of sampling error for a 50/50 result at 95% confidence is roughly:

Margin of error ≈ 1.96 × √(0.25 ÷ n)

Going from 1,000 to 2,000 completes doubles your sample cost but improves precision by less than a point. AAPOR makes the same point about diminishing returns and notes that subgroups have larger margins than the full sample. One important caveat from AAPOR: the margin of sampling error applies to probability-based samples, not to opt-in online panels. With opt-in samples, use these figures as a rough guide to how much your results would bounce around, not as a precise accuracy guarantee.

If you'll analyze subgroups (for example, comparing four US regions), size the study so each subgroup has enough completes. That, more than the total, usually sets the budget. The sample size calculator does this math for you.

3. Survey length

Longer surveys cost more per complete and lose more people partway through. Pew Research Center caps its own online panel surveys at 15 minutes and notes that the longer a survey is, the fewer people are willing to complete it, especially online. Pew budgets length using a point system (for example, a standalone question counts as 1 point, an open-ended question as 5) and treats 85 points as a 15-minute survey. You can borrow that approach to estimate your own length before you request a quote.

4. Audience difficulty

General-population US adults are the easiest audience to reach. Costs rise for:

• Narrow behaviors ("bought a used EV in the past six months")
• Professionals and decision-makers (IT buyers, physicians, small-business owners)
• Small geographic areas (a single metro area or county)
• Groups that are underrepresented on online panels

5. Quotas

Quotas (for example, matching US Census proportions by age, gender and region) improve balance, but each extra quota cell slows fielding. The last few completes in a hard cell are usually the most expensive. Quota only on variables that are likely to affect your results.

6. Data quality controls

Paying less for sample can cost more in the end if you have to throw out bad data. Pew Research Center found in a 2020 study that opt-in online sources contained roughly 4% to 7% bogus respondents, compared with about 1% in address-recruited panels. In a 2023 comparison, Pew found opt-in samples were less accurate than probability-based panels on average, with especially large errors for adults under 30 and Hispanic adults. Quality checks, replacement of rejected responses and manual review all have a cost, but skipping them risks drawing conclusions from noise.

7. DIY vs. agency

An agency's fee buys expertise and time. For a straightforward concept test or customer survey, many teams find DIY plus a quality-controlled panel is enough. For complex methods, the expertise is often worth paying for.

A budget formula you can use

Here is a simple formula to estimate a DIY market research survey:

Total budget = Sample + Platform + Your time + Contingency

Where:

• Sample = completes needed × price per complete (from your quote, which already reflects incidence, length and audience)
• Platform = monthly plan cost × months you'll need it
• Your time = hours for design, testing and analysis × your internal hourly cost
• Contingency = a buffer (many teams use 10–20%) for re-fielding, extra quota cells or scope changes

Worked example

Illustrative example only. The price per complete and hourly cost below are placeholder numbers chosen to make the math easy to follow; they are not SurveyHood prices or market benchmarks. Get a real price per complete from an instant quote.

You want to test three packaging concepts with US adults who buy plant-based snacks at least monthly. You plan to compare results across three age groups, and you want about 150 completes in each, so 450 total.

Now see how sensitive the total is to the drivers above. If you realize you only need to compare two age groups, you might need 300 completes instead of 450, cutting the sample line by a third. If, on the other hand, you narrow the audience to "buys plant-based snacks weekly," incidence drops and the price per complete in your quote will likely rise.

How to cut costs without hurting quality

Cut questions, not completes. Every question should map to a decision. Removing "nice to know" questions reduces length, which lowers cost and reduces drop-off.

Size the study to the decision. If you only need to know which of three concepts wins clearly, you may not need 1,000 completes. Use the sample size calculator to find the smallest sample that answers your question.

Use your own customers where they're the right audience. For satisfaction, onboarding and product feedback, your own list is cheaper and more relevant than a panel. Save paid sample for non-customers and market sizing.

Loosen non-essential criteria. If "bought in the past 6 months" and "past 12 months" would give you the same insight, the wider window raises incidence and lowers cost.

Simplify quotas. Quota on two or three variables that matter, and check the rest after fielding.

Put screeners first and keep them short. People who don't qualify should exit quickly.

Soft-launch. Field a small batch first, check the data, timing and drop-off points, then release the rest. Fixing a broken question after 50 responses is much cheaper than after 500.

Don't pay for bad data. Look for a sample source that removes speeders, straight-liners, out-of-country and duplicate respondents, and that only charges for responses that pass.

Avoid rework in analysis. Plan your cross-tabs before fielding so you collect exactly the variables you need to cut by.

What you get with SurveyHood

SurveyHood plans start free (3 surveys, 100 responses a month), with paid plans at $19, $49 and $99 a month and 20% off on annual billing; see pricing. If you need responses from outside your own list, you can buy survey responses: choose a country and audience (age, gender, region and profile attributes) and get an instant quote per approved complete before you pay. Responses go through automated checks and human review, you pay only for approved completes, and unused budget is refunded when the order closes. Results include live dashboards, cross-tabs and exports to CSV, Excel and PowerPoint. To start drafting, use the new product market research template or visit the market research survey page.

Budget checklist

1. Define the audience precisely and estimate incidence.
2. Decide which subgroups you'll compare and size each one.
3. Draft the questionnaire and estimate its length.
4. Get a quote for sample.
5. Add platform, internal time and a contingency.
6. Look for savings in length, criteria and quotas before cutting sample or quality checks.