Short answer: a traditional two-group in-person focus group project runs about $8,000 to $12,000 in a major metro area, and an online equivalent about $4,000 to $7,000, according to industry publication GreenBook. But the line item that actually determines your cost is not on any quote. It is the show rate - the share of recruited participants who turn up - and at the industry-average 80 percent you must recruit 13 people to reliably seat 10. For hard-to-reach professional audiences, where show rates collapse, the multiplier is what makes the project expensive, not the incentive.
This piece breaks down every component of focus group cost, shows you the arithmetic that vendors quote around rather than through, and gives you the one number to ask for before you sign anything.
What a focus group actually costs in 2026
There is no official price list for qualitative fieldwork, and the search results for this question are dominated by content marketing from firms that sell the service. Numbers vary by an order of magnitude between sources, which should tell you something on its own.
The most consistent published breakdown comes from GreenBook, the insights industry publication. Its figures are the publication's own estimates rather than a survey of transacted prices, so treat them as a planning range, not a benchmark:
| Component | In-person | Online |
|---|---|---|
| Facility rental | $1,500 - $2,500 per session | Platform license $500 - $1,500 per project |
| Participant incentives | $100 - $150 per person | $75 - $100 per person |
| Moderator and recruitment | Quoted separately | Quoted separately |
| Typical total project | $8,000 - $12,000 (two groups, major metro) | $4,000 - $7,000 |
GreenBook also lists the costs that never appear on the estimate: travel, lodging, catering for client observers, recording, staffing, transcription, parking, and car service for ethnographic work. Marta Villanueva, Chief Insights Officer at NuThinking, warns in the same piece about overages on platform hours and extended licenses for online sessions - the online equivalent of the catering bill.
For context on the scale of the industry doing this work: ESOMAR's Global Market Research 2025 report puts the insights industry above US$150 billion as of 2024, of which the traditional market research sector is US$56 billion, research software US$62 billion, and reporting US$35 billion. Notably, software has now overtaken services.
The line item nobody quotes: the over-recruit tax
Here is the arithmetic that decides your budget, and almost no proposal states it explicitly.
You do not pay for the participants you want. You pay for the participants you recruit, and you recruit more than you want because some will not show up.
Writing in Quirk's, Peter DePaulo assembled the best public data on this. The QRCA Field Committee consensus is that 80 percent is about an average show rate and 90 percent is a very good one. That consensus is backed by actual records: Merrill Shugoll examined 100 randomly selected groups at Shugoll Research and found 80.2 percent of recruits showed up. Anndel Martin polled 15 GroupNet facility members, who reported show rates between 75 and 85 percent.
DePaulo then does the statistics almost nobody does. To hold the risk of an under-attended group below 1 percent:
| Show rate | Recruits needed to seat 10 | Effective multiplier |
|---|---|---|
| 90 percent (very good) | 13 | 1.3x |
| 80 percent (average) | 17 | 1.7x |
| 50 percent (difficult population) | 27 | 2.7x |
Read the bottom row carefully, because it is where most B2B and healthcare research lives. To seat ten specialists you may need to recruit twenty-seven people.
This reframes what "expensive qual" means. The common explanation is that professional audiences cost more because the incentive is $500 instead of $100. That is true but secondary. The dominant term is the multiplier: a 5x incentive combined with a 2.7x recruitment multiplier is a 13.5x cost per seated participant, and only the first 5x appears on the quote you compare against.
There is a second, quieter cost on the same page. DePaulo notes that recruiting 12 for a 10-person group at an 80 percent show rate produces a 27.5 percent chance of over-shows - more people arriving than you can seat. Improve recruiting to an 85 percent show rate and the chance of over-shows rises to 44 percent. Those people are paid the full incentive and sent home. Tightening your recruiting does not remove the cost; it moves it from the risk of a failed group to the certainty of paying people who never speak.
The question to ask any vendor before you sign: not "what is the incentive?" but "what show rate are you assuming, how many are you recruiting, and who pays for over-shows?" A proposal that cannot answer all three is quoting you a price for a group it has not yet worked out how to fill.
Where the money actually goes
Grouping the components by who captures the spend clarifies which costs are negotiable and which are structural:
| Cost | Who captures it | Scales with | Negotiable? |
|---|---|---|---|
| Facility rental | Fieldwork agency | Sessions, city | Somewhat - off-peak, multi-group bookings |
| Incentives | Participants | Recruits, not attendees | Rarely - set by audience scarcity |
| Recruitment | Agency or panel | Screening difficulty, incidence | No - this is the real driver |
| Moderator | Independent or agency | Sessions and prep | Somewhat - experience tiers vary widely |
| Transcription and analysis | Agency, vendor, or you | Hours of audio | Yes - increasingly automatable |
| Client travel and observation | You | Locations | Yes - and the first thing to cut |
The pattern: the two largest and least negotiable items - incentives and recruitment - both scale with the number of people you must contact, not the number you actually learn from. Everything you pay is indexed to a quantity that has no relationship to how much insight you get.
That is the structural problem with the format, and it is worth stating plainly: a focus group's cost scales with logistics, while its value scales with the quality of the conversation. Those are two different variables, and the first one is the only one on the invoice.
Related reading on how the same arithmetic works for quantitative sample: How Much Does Survey Sample Cost in 2026? CPI, Incidence Rate and LOI Explained. Incidence rate there plays exactly the role show rate plays here.
The costs that are not money
Two more costs belong in any honest comparison, because they routinely exceed the invoice.
Time. A traditional focus group project runs on a calendar measured in weeks: write the screener, field the recruit, wait for the facility slot, run the sessions, transcribe, analyze, write up. If your decision deadline is inside three weeks, the method is unavailable at any price. Our user research cost calculator walks through the full timeline alongside the budget.
Group effects. Focus groups have a well-documented methodological cost: dominant participants set the frame, quiet participants converge on it, and moderator behaviour shapes what gets said. We cover this in depth in Focus Groups: Advantages and Disadvantages and Focus Groups vs Interviews. The short version is that you are paying premium logistics costs for a format that structurally suppresses minority views - and minority views are usually the reason you commissioned research.
If you are running groups anyway, the practical guides are How to Conduct a Focus Group, the focus group script template, focus group questions, and how to analyze focus group data. Getting the screener right is the single highest-leverage cost control available - see screener questions and research screener questions.
What changes the arithmetic
The over-recruit tax exists because a focus group is an appointment: a fixed number of people must be in one place at one time, and every absence damages the session. Any method that removes the appointment removes the multiplier.
AI-moderated interviews do exactly that. Because each conversation is independent and asynchronous, a no-show is not a failure mode - it is simply a conversation that has not happened yet. There is no facility standing empty, no eight other participants waiting, no over-show to pay off. You do not recruit 17 to seat 10; you invite people and count completed conversations.
The cost structure changes shape entirely. On Koji, interviews start as low as €1 per qualified interview and €3 per qualified voice interview, and credits are the only unit. Start with pay as you go. No subscription needed. You pay only for the interviews your study actually uses. A conversation that scores below 3 out of 5 is free and never charged, so low-effort sessions do not bill. Volume pricing and plans are there when you want them.
Set that against the GreenBook range honestly. A $8,000 two-group project puts roughly 16 to 20 people in rooms. The comparison is not perfectly like-for-like: an agency project bundles recruitment, a professional moderator, facility logistics, and a written deliverable, and if you need a national probability sample you still need a sample provider. But on the thing both methods actually produce - a recorded, analyzed conversation with a real customer - the per-conversation cost differs by roughly two orders of magnitude, and the calendar differs by weeks.
The depth objection deserves a straight answer too. AI moderation does not get tired, does not lead the witness, and does not run out of patience on the fourteenth interview - the failure modes that make human moderation expensive to do well and inconsistent to do at scale. Koji's interviews probe follow-ups automatically on open-ended answers, and support six structured question types alongside them - open_ended, scale, single_choice, multiple_choice, ranking, and yes_no - so a single study returns both quotable narrative and countable distributions. That is the combination a focus group cannot give you: groups produce transcript, not tabulation. See structured questions for how the two halves fit together.
How to budget a qualitative project in 2026
A defensible process, in order:
- Define the audience precisely, then estimate incidence. The share of the population that qualifies drives recruitment cost more than any other input.
- Ask the vendor for the assumed show rate in writing. Then compute your own multiplier from the table above. If they assume 90 percent for a physician audience, they are quoting a project they cannot deliver.
- Price the recruit, not the seat. Multiply incentive by recruits, not by attendees, and ask explicitly who absorbs over-shows.
- Add the invisible line items. Transcription, observer travel, catering, platform overage.
- Put a fast asynchronous method in the same comparison. Not to replace groups everywhere, but because the price of the alternative is the only thing that makes the group's price legible.
- Decide what you would do differently at each possible finding. If no finding changes a decision, the cheapest correct budget is zero.
For teams working under real constraints, user research on a budget and the user research budget template are the practical companions to this piece. If you are weighing vendors rather than methods, see best market research companies and participant recruitment platforms.
Where Koji fits
Koji is an AI-native customer research platform built for teams who need the depth of a qualitative conversation without the logistics that make qualitative expensive.
- AI-moderated voice and text interviews that run asynchronously, so there is no appointment to fill and no over-recruit tax.
- Automatic thematic analysis across every transcript, so the analysis stage does not scale linearly with hours of audio.
- Six structured question types alongside open-ended probing, so one study yields both quotes and distributions.
- Customizable AI consultants that carry your brief and probe the way your best moderator would - consistently, on the first interview and the hundredth.
- One-click reports you can hand to a stakeholder the same day the last interview lands.
- No moderator bias and no group effects - every participant answers in their own words without a dominant voice in the room.
Legacy providers - UserTesting, Qualtrics, Dovetail, dscout, and the traditional fieldwork agencies - price research as a logistics problem because for them it is one. Koji prices it as a conversation, which is what you were trying to buy.
Ready to find out what your next study actually costs? Start your first Koji study free and get from question to insight in hours, not weeks - with no research expertise required.
Frequently Asked Questions
How much does a focus group cost in 2026?
Industry publication GreenBook puts a traditional two-group in-person project at $8,000 to $12,000 in a major metro area, and an online project at $4,000 to $7,000. Facility rental runs $1,500 to $2,500 per session in person, with online platform licenses at $500 to $1,500 per project. Incentives are $100 to $150 per person in person and $75 to $100 online. Moderator and recruitment fees are quoted separately and vary widely by audience difficulty.
Why do focus groups with professionals cost so much more?
Two multipliers stack. Specialist incentives are several times higher than consumer incentives, and specialist show rates are far lower. At a 50 percent show rate you must recruit 27 people to reliably seat 10, versus 13 at a 90 percent show rate. A 5x incentive combined with a 2.7x recruitment multiplier produces roughly 13.5x the cost per seated participant, and only the incentive difference typically appears on the quote.
What is a normal focus group show rate?
The QRCA Field Committee consensus reported in Quirk's is that 80 percent is about average and 90 percent is very good. Merrill Shugoll's review of 100 randomly selected groups at Shugoll Research found 80.2 percent, and a poll of 15 GroupNet facility members reported 75 to 85 percent. Anything near 50 percent indicates a genuinely difficult population and should change your recruitment budget substantially.
How many people should I recruit for a focus group?
To keep the risk of an under-attended group below 1 percent, recruit 13 for a 10-person group at a 90 percent show rate, 17 at 80 percent, and 27 at 50 percent. Be aware of the trade-off: over-recruiting to protect the session raises the chance of over-shows, which reach roughly 44 percent at an 85 percent show rate. Those participants receive full incentives without contributing data.
Are online focus groups cheaper than in-person?
Yes, typically by about half - GreenBook's ranges are $4,000 to $7,000 online versus $8,000 to $12,000 in person. Online removes facility rental, travel, and catering, and digital incentives are somewhat lower. It does not remove the recruitment multiplier, because the session is still a scheduled appointment that a fixed number of people must attend simultaneously. Watch for platform hour overages and extended license fees.
What is the cheapest way to get qualitative insight?
Removing the appointment removes most of the cost. Asynchronous AI-moderated interviews eliminate facility rental, scheduling, over-recruitment, and over-show payments, because each conversation is independent. Koji interviews start as low as €1 per qualified interview and €3 per qualified voice interview. Start with pay as you go. No subscription needed. You pay only for the interviews your study actually uses. Traditional groups remain useful when you specifically need participants to react to each other.
