# Focus Group Research: Methodology, Online Platforms & Interviewing

> How to run focus group research: group size, recruiting, discussion guide, moderation, analysis, online formats and platforms, with limits and FAQs.

Source: https://www.pulselake.co/blog/focus-group-research-methodology-online-platforms-2026
By PulseLake Team · PulseLake

A focus group is a planned discussion in which a trained moderator guides roughly 6 to 10 purposively recruited people through semi-structured questions, and the interaction between them is the main data. It's a qualitative method: it explains why people hold views and how they talk about them, not how many share them. The method runs in six steps: define the question, recruit with a screener, write a discussion guide, moderate, record and transcribe, then code themes across groups.

We make PulseLake, a research platform. It comes up once in this guide, near the end, and we say plainly where it doesn't apply.

![Moderator and six participants around a table with a discussion guide, then record, transcribe and find themes.](https://www.pulselake.co/blog/img/production/3765b2a1176122f86b1b87a958bc5b96aed9cba6-1200x840.png?w=1600&fit=max&auto=format)

## What is focus group research?

It's how you hear a target audience's attitudes, beliefs, perceptions and vocabulary through facilitated group conversation. MAXQDA's research guide describes group interaction as the primary data source and lists typical uses: exploring perceptions, reacting to concepts, ads or packaging before quantitative testing, and explaining the reasons behind survey results ([MAXQDA](https://maxqda.com/research-guides/focus-group-research)).

Samples are small and purposive, so the findings can't be generalized statistically to a population ([MAXQDA](https://maxqda.com/research-guides/focus-group-research)). Need a prevalence number, like "what share of buyers prefer option B?" Run a survey. Our piece on [qualitative vs quantitative research](https://www.pulselake.co/blog/qualitative-vs-quantitative-beyond-the-basics) covers how the two fit together.

## When should you use a focus group instead of interviews or a survey?

| Method | Use it when | Weak at |
| --- | --- | --- |
| Focus group | Group reactions and shared language matter | Sensitive topics, prevalence |
| In-depth interview | You need one person's detailed story | Seeing group dynamics |
| Survey | You need counts and comparisons | Explaining the "why" |

MAXQDA gives group synergy yielding richer data as the reason to prefer a group over individual interviews, and cautions against groups where authority relationships among participants may produce conformity ([MAXQDA](https://maxqda.com/research-guides/focus-group-research)). The classic one to avoid is a group of employees with their manager in the room.

## What does focus group methodology look like step by step?

### 1. Define the research question

Before anything else, settle which decision the findings will inform. MAXQDA suggests three to five research questions, which stops a single session from sprawling ([MAXQDA](https://maxqda.com/research-guides/focus-group-research)). Our guide to [defining research problems and objectives](https://www.pulselake.co/blog/defining-research-problems-and-objectives) helps with this step.

### 2. Plan the number and size of groups

Published ranges differ, so take them as starting points, not rules:

- Great Question gives 6 to 10 participants as standard, 4 or 5 for "mini groups" on sensitive or complex topics, and three to six groups per audience segment, stopping when new groups stop producing new insight ([Great Question](https://www.greatquestion.com/ux-research/focus-groups-guide)).
- MAXQDA reports 6 to 10 as typical, 3 to 5 for complex topics and up to 12 for simple ones, and gives 3 to 4 groups per segment as a common starting point, noting that two can suffice with well-composed groups and a skilled moderator ([MAXQDA](https://maxqda.com/research-guides/focus-group-research)).

What they share: plan several groups per segment, and use saturation (new groups add little that's new) as your signal to stop.

### 3. Recruit with a screener

A screener is a short questionnaire that checks candidates against your criteria before you give them a seat. MAXQDA suggests 5 to 15 questions ([MAXQDA](https://maxqda.com/research-guides/focus-group-research)). Recruit people who share the trait relevant to your question, and keep groups fairly homogeneous so people feel comfortable disagreeing. See [avoiding sampling bias](https://www.pulselake.co/blog/avoiding-sampling-bias) for how recruitment choices skew results.

Over-recruit to cover no-shows. Great Question suggests 20% to 30% more people than you need and a confirmation 24 to 48 hours ahead ([Great Question](https://www.greatquestion.com/ux-research/focus-groups-guide)); MAXQDA says 20% and 48 hours ([MAXQDA](https://maxqda.com/research-guides/focus-group-research)).

**Hypothetical example (invented numbers):** you want 8 people per group across 3 groups. At 25% over-recruitment, that's 8 x 1.25 = 10 confirmed recruits per group, so 10 x 3 = 30 recruits for 24 seats.

### 4. Write the discussion guide

Think of the guide as a roadmap. It isn't a script. Great Question recommends open-ended prompts such as "Walk me through how you...", moving from broad to specific, and leaving time for follow-ups on unexpected points ([Great Question](https://www.greatquestion.com/ux-research/focus-groups-guide)). MAXQDA suggests three to five key questions ([MAXQDA](https://maxqda.com/research-guides/focus-group-research)). A typical shape (our summary):

1. Welcome, ground rules, consent to record.
1. Warm-up: an easy question every person answers.
1. Core topics, broad first, then specific.
1. Stimulus: a concept, ad or prototype ([concept testing guide](https://www.pulselake.co/blog/concept-testing-guide-methods-questionnaire-examples-2026) and [message testing](https://www.pulselake.co/blog/marketing-message-testing-methods-survey-questions-2026) cover this).
1. Closing round: "Is there anything we did not ask that we should have?"

Steer clear of leading wording. Everything in [writing questions that don't bias responses](https://www.pulselake.co/blog/writing-questions-that-dont-bias-responses) applies to a discussion guide too.

### 5. Moderate the session

A good moderator guides without dominating. Great Question advises setting ground rules such as one speaker at a time, using neutral follow-ups such as "Can you tell me more?", avoiding praise or agreement with particular answers, drawing out quieter people, redirecting dominant ones, and watching for groupthink ([Great Question](https://www.greatquestion.com/ux-research/focus-groups-guide)). Some practical habits:

- Ask the group to hold a question for later instead of cutting a speaker off.
- Go round the table on key questions so everyone commits to an answer before hearing the others.
- Dig into disagreement. Don't smooth it over.

That last one is our strongest opinion here. The disagreements are usually where the insight is.

### 6. Record, transcribe and analyze

Get recording consent in writing beforehand and verbally at the start ([Great Question](https://www.greatquestion.com/ux-research/focus-groups-guide)). MAXQDA recommends a hot debrief within 30 minutes of the session ending, while impressions are still fresh ([MAXQDA](https://maxqda.com/research-guides/focus-group-research)).

For analysis, Great Question suggests transcribing, reviewing several times, tagging comments with themes, noting where opinions differ, backing each finding with quotes and checking back against the original objective ([Great Question](https://www.greatquestion.com/ux-research/focus-groups-guide)). MAXQDA adds that deductive (predefined) and inductive (emergent) coding are typically combined, and advises against analyzing individual opinions in isolation from group dynamics ([MAXQDA](https://maxqda.com/research-guides/focus-group-research)). Compare themes across groups as well as within them. For larger transcript sets, see [qualitative coding at scale](https://www.pulselake.co/blog/qualitative-coding-at-scale) and [AI-assisted theme extraction](https://www.pulselake.co/blog/ai-assisted-theme-extraction); a human should still review the codes.

## What are the limitations of focus groups?

- Groupthink, where people converge on a shared view out of politeness or because one voice dominates.
- Social desirability, where people give answers that look good in front of others.
- Moderator influence on tone and direction.
- Harder analysis, because interaction data is more complex than individual answers.
- Recruiting effort, since getting the right people in one place at one time is hard.

MAXQDA lists all five ([MAXQDA](https://maxqda.com/research-guides/focus-group-research)). The fixes are mostly design choices: homogeneous groups, a skilled moderator, written pre-work before the session, and follow-up interviews for sensitive subjects.

## What are online focus groups?

An online focus group is a qualitative session held over the internet with dedicated software ([QuestionPro](https://www.questionpro.com/blog/online-focus-group/)). It comes in two formats.

|  | Synchronous | Asynchronous |
| --- | --- | --- |
| Timing | Live, at a set time | Respond when convenient |
| Length | 30 to 90 minutes (QuestionPro) | Days to weeks or longer |
| Size | 6 to 10 (QuestionPro) | 15 to 20 start (iTracks) |
| Suits | Concept and product reactions | Diaries, co-creation, tracking |

QuestionPro puts synchronous sessions at 30 to 90 minutes with 6 to 10 participants, and describes asynchronous (bulletin board) groups as running over days, weeks or even years ([QuestionPro](https://www.questionpro.com/blog/online-focus-group/)). MAXQDA gives 60 to 90 minutes for online synchronous sessions and 2 to 5 days for asynchronous ([MAXQDA](https://maxqda.com/research-guides/focus-group-research)). iTracks says the most common asynchronous run is 2 to 3 days, with 15 to 20 participants per group as a starting point, and says larger groups are easier to manage because participants don't all speak at once ([iTracks](https://www.itracks.com/blog/intro-to-asynch)).

Great Question's take on online versus in person: online is more convenient, cheaper and easier to schedule, and suits dispersed audiences and topics that don't involve handling a physical product. In person gives richer nonverbal cues and suits physical product testing ([Great Question](https://www.greatquestion.com/ux-research/focus-groups-guide)).

## Which online focus group platforms are there?

As of October 2026, here's what three vendors say on their own public pages. It isn't a ranking, and we give no prices. The "Our read" lines summarize each page.

| Platform | What its page says | Our read |
| --- | --- | --- |
| [iTracks](https://www.itracks.com/) | Live video focus groups and interviews, asynchronous text, and discussion boards, diaries and communities over multiple days | Covers live and asynchronous in one offering |
| [Recollective](https://www.recollective.com/) | Online focus groups, bulletin-board studies, diaries and ethnography; live video plus self-paced tasks | Mixes live and asynchronous work |
| [Discuss.io](https://www.discuss.io/) | Focus groups and IDIs in more than 100 markets, with an observer room and live translation into English | Live sessions across many markets |

How to choose? Decide synchronous or asynchronous first. Then check recruitment support, observer access, recording and transcription, language needs and data handling. A general video-call tool can work for a small project, but it won't manage recruitment or analysis for you.

## How does AI change moderation?

Some platforms now offer AI-moderated interviews, which run one-to-one conversations instead of live groups. That's a different method with different trade-offs, covered in our guide to [AI focus groups and AI-moderated interviews](https://www.pulselake.co/blog/ai-focus-groups-and-ai-moderated-interviews-guide-2026) and the short answer on [what an AI-moderated interview is](https://www.pulselake.co/answers/what-is-an-ai-moderated-interview). This page sticks to human-moderated groups.

Where PulseLake fits: its public materials describe traditional research including qualitative research, and AI-led interviews with real participants. This guide doesn't claim PulseLake runs live, human-moderated group video sessions. For those, use a dedicated platform like the ones above.

## Frequently asked questions

### How many people should be in a focus group?

Six to 10 is what most sources give. Great Question suggests 4 or 5 for sensitive or complex topics, MAXQDA says 3 to 5 for complex topics, and MAXQDA allows up to 12 for simple ones.

### How many focus groups do you need?

Plan several per audience segment. Great Question suggests three to six per segment and stopping when new groups add little; MAXQDA gives 3 to 4 per segment as a common starting point and says two can suffice with well-composed groups and a skilled moderator.

### How long should a focus group last?

MAXQDA gives 60 to 120 minutes in person and 60 to 90 minutes for live online sessions. QuestionPro puts synchronous online sessions at 30 to 90 minutes.

### What is the difference between a focus group and an in-depth interview?

A focus group studies how people discuss a topic together and treats their interaction as data. An in-depth interview explores one person's experience. Pick interviews for sensitive topics, or whenever group pressure would distort answers.

### Are online focus groups as good as in-person ones?

They're good at different jobs. Online is cheaper and easier to schedule for dispersed audiences, while in person gives richer nonverbal cues and suits physical products ([Great Question](https://www.greatquestion.com/ux-research/focus-groups-guide)).

### Can focus group results be generalized?

No. Samples are small and purposive, so findings can't be statistically generalized ([MAXQDA](https://maxqda.com/research-guides/focus-group-research)). Use them to form hypotheses, then test those with a survey.

## Related reading

- [AI Focus Groups and AI-Moderated Interviews: The Complete Guide](https://www.pulselake.co/blog/ai-focus-groups-and-ai-moderated-interviews-guide-2026)
- [Customer Research Methods: Modern Techniques for Insights Teams (2026)](https://www.pulselake.co/blog/customer-research-methods-2026)
- [Qualitative vs Quantitative Research: Beyond the Basics](https://www.pulselake.co/blog/qualitative-vs-quantitative-beyond-the-basics)
