The Choice Comes Down to Individual Reasoning or Group Reaction

An in-depth interview, usually shortened to IDI in market research, is a one-on-one conversation between a trained interviewer and a single participant, typically running 45 to 90 minutes. The interviewer follows a guide but has room to chase anything worth following, and the result is a detailed account of how one person reached their views.

A focus group brings six to ten participants together with a moderator for roughly 90 minutes to two hours. The moderator still works from a guide, but much of the value comes from what happens between participants as they build on each other's comments and push back. That interaction is data in its own right, and an interview has no way to produce it.

Most method decisions follow from this. A question about a personal decision or expert judgment needs the individual account, and a question about how an idea holds up once other people start talking about it needs the group. The guide on how to conduct a focus group covers running a session.

Each Method Gives Up Something the Other Keeps

Interviews give up group energy in return for depth and control, and focus groups make the opposite trade.

Start With the Research Question, Not the Method

Teams that pick the method first often bend the question to fit it. Working backward from the question avoids that, and the table below shows how it plays out for common research briefs.

Across market, UX, healthcare, and academic research, questions that start with "why did you" usually point to interviews, and questions that start with "how do people react to" usually point to groups.

Dyads, Triads, and Mini-Groups Fill the Space Between

Paired interviews, or dyads, put two people who share a decision in one conversation, such as a couple or two members of a buying committee, and triads add a third voice when the goal is to watch a small disagreement play out without the crowd effects of a full group. Mini-groups of four to six give each person more airtime than a full focus group. They work well for specialists who are hard to recruit in numbers, and for online sessions, where eight video tiles get difficult to moderate.

Interviews Reach Saturation for Less Than Most Teams Assume

The most useful evidence on this question comes from two papers by researchers at FHI 360. In the first, Namey, Guest, McKenna, and Chen (2016, American Journal of Evaluation) measured how many sessions each method needed to reach thematic saturation, the point where new sessions stop producing new themes. In their data, 80 percent saturation took eight interviews or two to four focus groups, and 90 percent saturation took sixteen interviews or five focus groups. Once the authors costed each approach, including moderator time, transcription, and participant incentives, individual interviews came out 12 to 20 percent cheaper than focus groups, and that result held when they varied those cost assumptions.

The second paper, Guest, Namey, Taylor, Eley, and McKenna (2017), published in the International Journal of Social Research Methodology, randomly assigned 350 African American men in Durham, North Carolina, to either interviews or focus groups covering the same questions about seeking health care. Session for session, both generated a similar number of unique items, but per participant, interviews were more effective at generating a broad range of items.

These numbers come from one population and one topic, so they are a starting point rather than a rule, but they do challenge the common assumption that focus groups are the efficient choice because they cover more people per session.

Sensitive Topics Don't Automatically Belong in One-on-One Interviews

The usual advice is to keep sensitive subjects out of groups, because people may not want to talk about their health or money in front of strangers. The Guest randomized study complicates that advice, finding that several types of sensitive and personal disclosures were more likely in focus groups, some sensitive themes appeared only in the group sessions, and no sensitive theme came up more often in the individual interviews.

Hearing someone else go first may make disclosure feel normal in a way a one-on-one setting cannot, but that does not make groups right for every sensitive topic. Interviews are still the safer choice when a disclosure could identify someone in their community or when participants know each other outside the session. Groups also have a confidentiality limit interviews do not, since a researcher can ask participants to keep what they hear private but cannot guarantee it.

Each Method Brings Its Own Kind of Bias

Focus groups are vulnerable to conformity, where participants drift toward an early or confident opinion, and to dominant voices that take over the session. Experienced moderators collect written reactions before opening the discussion and treat early consensus with some suspicion until the quieter participants have spoken. A deliberately varied recruit helps as well, as the guide to diversity in focus group participants explains.

Interview bias is harder to spot. The interviewer's tone and choice of follow-up questions shape what the participant says next, and with nobody else in the room, a tidy but incomplete account goes unchallenged. A discussion guide with neutral probes protects both formats, as does checking early transcripts for leading questions while there is still time to adjust. Teams considering automated moderation should read the comparison of AI and human moderators first.

Focus Groups Earn Their Cost When the Interaction Is the Data

Groups are worth the extra recruitment and incentive costs when the research depends on people responding to each other. Concept and message testing fall into that category, and so do co-creation sessions where participants develop ideas together.

Most projects run at least two groups per audience segment so that one unusual session does not drive the findings, and the Namey data points to three to five groups per segment for most of the themes in a population. Interview studies commonly plan 12 to 20 interviews per segment and adjust once new themes taper off, and in a study with several segments, those numbers multiply fast enough for the cost difference between methods to matter.

Strong Studies Often Use Both Methods

A sequential design might open with two or three focus groups to map the range of opinions and the language people use, then follow with interviews that dig into the most important threads individually. The reverse order works too, with interviews uncovering individual experiences and groups testing whether those experiences resonate more widely, and both pair well with survey work, as the post on mixed methods in market research describes. Mixed designs do require speaker labels, participant IDs, and formatting that match across interviews and groups, so the whole dataset can be coded together in NVivo, ATLAS.ti, or MAXQDA.

The Method You Choose Decides What Your Transcript Has to Handle

An interview transcript has two speakers taking clean turns, and its main demands are accuracy, the right verbatim style, and careful treatment of names and terminology. A focus group transcript has to follow up to ten voices through interruptions and moments when three people answer at once, and it has to keep every comment attached to the right person. When attribution slips, a point made repeatedly by one participant can look like group consensus during coding, and nothing later in the analysis is likely to catch it.

Qualtranscribe's human focus group transcription is built for that second case, with consistent moderator and participant labels through overlapping talk, timestamps, and a choice of clean or full verbatim (the verbatim guide explains which suits which kind of analysis). Interview transcription follows the same standard for one-on-one and paired sessions, and both arrive formatted for NVivo, ATLAS.ti, and MAXQDA.

Remote sessions bring uneven microphones, connection drops, and people talking over a lag, and Qualtranscribe transcribes Zoom, Teams, and Webex recordings with the same speaker identification it applies in person (the guide to running a focus group on Zoom covers recording settings that help). Human transcription and direct translation to English are available in 25 languages, so a study with groups in Spanish and interviews in Japanese still ends up as one dataset.

When the moderator wants to adjust the guide before the next session, Instant Draft produces an AI transcript in minutes and Smart Insights surfaces themes and key quotes linked by timestamp to the recording, while human transcription produces the final record for coding and client delivery. The guides on analyzing interview transcripts and moving from transcripts to themes take it from there.

What Transcription Costs Look Like for the Same Study Run Both Ways

The table below prices one research question both ways at the saturation levels from the Namey data, using Qualtranscribe's English human transcription rates of $1.20 per minute for interviews and $1.70 per minute for focus groups (the standard rate plus $0.50 for multi-speaker sessions). Session lengths of 60 minutes per interview and 90 minutes per group are assumptions for illustration, and the 80 percent group figure uses three groups, the midpoint of the two-to-four range.

On transcription alone the focus group design costs less, because a group packs several participants into each recorded minute, but transcription is usually a small share of either budget next to recruitment, incentives for every group participant, facility or platform fees, and moderator time. That is why Namey and colleagues found interviews cheaper once the full study was counted, and why teams comparing designs should price the whole study. The pricing page lists per-minute rates for languages beyond English.

Group Sessions Change Your Privacy Obligations

Every participant in a focus group needs to consent to being recorded, and the consent form should explain that other participants will hear what they say. Moderators usually ask participants to use first names or pseudonyms and to keep the discussion private afterward, even though that request cannot be enforced.

A group transcript can also carry identifying details for eight or more people. Each one needs a consistent replacement, or the coding structure falls apart. Qualtranscribe offers participant de-identification on request for human transcription, and the explainer on de-identification, anonymization, and pseudonymization sets out which approach a protocol may require.

Which regulations apply depends on who is in the room: patient focus groups and clinical interviews in the US fall under HIPAA, participants in the EU fall under GDPR, Canadian research falls under PIPEDA, and studies with Japanese participants fall under APPI. Qualtranscribe handles every file under NDA and keeps recordings encrypted on regionally isolated servers. Recordings are never used to train AI models. Teams that need no AI processing of protected health information at all can choose human transcription only.

If your next study involves interviews, focus groups, or both, you can upload your first recording here for human transcription or an Instant Draft.

Frequently Asked Questions

What is an IDI in market research? IDI stands for in-depth interview, a one-on-one qualitative conversation between a trained interviewer and a single participant, usually lasting 45 to 90 minutes. Market researchers use IDIs for B2B buyers, healthcare professionals, and any topic where individual decision-making matters more than group reaction.

How many in-depth interviews do you need? It depends on how varied the population is, but in the Namey et al. (2016) data, eight interviews reached 80 percent thematic saturation and sixteen reached 90 percent within a single population. Most studies plan for 12 to 20 per segment and adjust once new themes stop appearing.

How many people should be in a focus group? Full focus groups usually have six to ten participants. Mini-groups of four to six give each person more time to talk and suit specialist audiences, product testing, and online sessions.

Are focus groups cheaper than in-depth interviews? Per session, focus groups cover more people, and they often cost less to transcribe because several participants share each recorded minute. Across a whole study, the evidence leans the other way, since Namey et al. found interviews reached saturation for 12 to 20 percent less once moderator, transcription, and incentive costs were counted.

Should sensitive topics always use in-depth interviews? Not necessarily, since a randomized study by Guest et al. (2017) found several kinds of sensitive disclosures were more likely in focus groups. Interviews remain the better choice when a disclosure could identify someone, when participants know each other, or when the topic carries legal risk.

Does focus group transcription cost more than interview transcription? At Qualtranscribe, English human transcription is $1.20 per minute, and focus groups are $1.70 per minute because multi-speaker sessions take more work to attribute accurately. Speaker labels, timestamps, and formatting for qualitative analysis software are included in both.