How to Choose a Qualitative Research Agency in India

A qualitative research agency may recruit, moderate and analyse in-house, use a named partner for one step, or subcontract steps study by study. None of these models is wrong. What hurts a study is not knowing which one you are buying, because the party you never meet decides who sits in your sessions.

Subcontracting is acceptable when you are told about it. Get it in writing and keep the agency you sign with answerable for every step. If the agency holds a certificate under ISO (the International Organization for Standardization) 20252, the standard for market, opinion and social research, ask which edition it covers. ISO 20252:2026 is the current edition, and a certificate may be to an earlier one. Ask also whether the services it covers include subcontracted work.

Who will run your project day to day?

Pass: a named project lead, a stated rhythm for updates (for example a call when recruitment is signed off, after the pilot and at the first topline) and a named senior person you can escalate to. Fail: an account manager you meet only at the pitch, or updates “as needed”.

Do you need an agency, a recruitment partner or a platform?

An agency designs the study, runs it and interprets it. A recruitment partner supplies participants and fieldwork support to a study that you or your own agency run. A platform is software you operate yourself. If you have researchers and need the right people in the room, a recruitment partner may be enough. If you need the question framed and answered, you need an agency. Our guide to platforms matched to the method covers the software route.

Write a one-page brief covering the decision the research will inform, the method, who is in and who is out, cities, language, sample, deliverables and timeline, and send the same brief to every agency. Quotes can only be compared when agencies priced the same thing.

What should your brief cover?

Start from the decision, not the method. Ask each agency to restate your decision in its own words. An agency that proposes a different method should explain why. Use a one-page brief template if you do not have one. Ask each agency to justify the sample from your research question and to say how it will know when it has heard enough. Our guide to how many interviews or groups to plan shows what a defensible answer looks like.

What should you check for each qualitative method?

In-depth interviews (IDIs) are one-to-one conversations. Focus group discussions (FGDs) are moderated group conversations. An agency's strength in one does not carry over to the others, so check each method on its own terms.

Check recruitment first, because a skilled moderator cannot rescue a room of the wrong people. Ask to see the screener, how identity is verified before invitation, how professional respondents are kept out, and which cities the recruiter actually reaches. For deeper criteria, see our guide to assessing participant panel quality.

Can you see the screener and the recruits' answers before fieldwork?

Pass: the agency shares the draft screener for your sign-off and gives you each recruit's answers, without names, before sessions. Fail: it refuses, or says eligibility is "confirmed by phone" and keeps no record.

How is identity checked before anyone is invited?

Pass: the agency describes a step you can repeat back, covering what is checked, by whom, and what happens to a recruit who fails. Ask what it does when it sees the fraud signals researchers report: surges of similar emails, mismatched contact or personal details, geographical inconsistencies and implausible answers.

Those signals come from a 2024 survey of 37 Canadian health care researchers, all of whom had suspected fraudulent participants in an online-recruited study. Of the 37, 31 (84%) said their recruitment messages had mentioned incentives (Kumarasamy and colleagues, JMIR Formative Research, 2024; JMIR stands for Journal of Medical Internet Research). The study cannot show how common fraud is, so our inference is limited: paid studies recruited online in India face the same incentive, which is why identity checks matter.

Fail: "we trust our database."

How are professional respondents kept out?

A professional respondent is someone who joins studies regularly for the incentive. Pass: a written rule with a look-back window and a cap on recent studies, checked against the agency's own records. Our recommendation is to exclude anyone who took part in a study in your category in the last six months, adjusted for how narrow your audience is. Pass also means a recruit who fails in the first minutes is replaced from the same quota cell (the target-group slot a recruit fills) with identical screening, and you are told. Fail: no records of past participation, or replacements taken from whoever is available.

Which cities can the recruiter reach, and through whom?

Pass: a city list with a named recruiter or partner for each. Fail: "pan-India" with no list. If you want to recruit some participants yourself, our guide to recruiting participants in India covers that route.

Does the agency check feasibility before it quotes?

Pass: before quoting, the agency tells you whether your audience can be found in your cities, how long recruitment will take by city tier, and how many extra recruits it will line up. IMARC Group’s 2026 guide to focus groups in India says a 15–25% no-show rate is standard even with confirmed participants and advises over-recruiting by 20–25% (IMARC Group, 2026).

Fail: an instant quote with no questions about how rare your audience is, or no plan for no-shows.

Judge moderation on three things you can see before you pay: the moderator, the language and the discussion guide. Speak to the moderator who will run your sessions, confirm the language participants will speak, and have the moderator review your discussion guide with you.

Can you speak to the moderator who will run your sessions?

Pass: a short call in which the moderator asks about your decision and challenges draft questions. Fail: no named moderator, or a different seniority from the pitch team.

For sensitive topics or mixed groups, ask whether the moderator’s gender and regional background suit the participants. Pass: the agency explains its choice. Fail: the moderator is allocated without reference to the audience.

Will sessions run in participants' own language, and who writes the transcript?

Pass: the moderator is fluent in the language of the session and you receive the original-language transcript alongside any English translation, so a bilingual colleague can check a page against it. Fail: "we cover all languages" with no moderator named per language, or transcripts delivered only as translation. For consumer work, our recommendation is to run sessions in the language participants are most comfortable in, and to record that language in the screener.

Ask also where AI tools are used, for example in transcription, translation or coding, and whether a researcher checks the output before you see it. Pass: each use is named and a person reviews it. Fail: “our platform handles it”, with no reviewer.

Will the moderator review your discussion guide?

The discussion guide is the planned list of topics and questions. Pass: you get a marked-up guide with the changes explained before fieldwork, such as leading questions rewritten. Fail: the guide is used as sent.

An agency should show you the professional code it works under, a sample consent form and a written account of what happens to recordings. The ICC/ESOMAR code, issued by the International Chamber of Commerce (ICC) and ESOMAR (the global market research association), and the Market Research Society of India (MRSI) code set professional standards. The DPDP Act adds legal duties.

The DPDP Rules were notified on 13 November 2025 (G.S.R. 846(E), published by the Ministry of Electronics and Information Technology). Core obligations, including the notice duty (Rule 3), security safeguards, breach intimation and rights of data principals (the individuals whose data is processed), apply from 13 May 2027 (as of 22 October 2026). Rule 14 requires data fiduciaries, the parties that decide why and how personal data is processed, to publish how people can make access, correction and erasure requests, and sets a maximum of 90 days for responding to grievances (Rule 14 applies from 13 May 2027).

In a study, the data fiduciary can be you, the agency or both. Settle it in the contract and have your legal adviser confirm it.

Three tests:

• Consent form. Pass: a sample form shows purpose, payment, recording, who sees the data, how long it is kept and how to withdraw, with separate choices for recording, client viewing and quotes. Our informed consent form and the Indian rules guide gives a model. Fail: "participants sign our standard form" and you do not see it.
• Recordings. Pass: the agency names who can access recordings, where they are stored, the retention period and the deletion process. Fail: no retention period.
• DPDP preparation. Ask how the agency is preparing. Pass: a named person or adviser and a written plan. Fail: "fully compliant" with no description, or "not relevant to us".

A comparable quote lists line items against your brief and names every deliverable before the study starts. Without that, two quotes cannot be compared, because one may include incentives, translation and revision rounds that the other leaves out.

We have no fixed rate card and price each project from its brief (as of October 2026). In our view, the lines that most often move a qualitative quote are recruitment difficulty, participant incentives, and the number of sessions and languages. Our recruitment cost worksheet and guide to what to pay focus group participants show how to budget them.

Ask every agency to price the same line items:

1. Recruitment and screening, including replacements
2. Participant incentives, and whether they sit inside the quote
3. Moderation, with the number of moderators and sessions each
4. Venue or online platform
5. Interpretation and translation
6. Recording and transcription, in the original language and translated
7. Analysis and report, with the number of revision rounds
8. Project management and travel

For deliverables (recordings, transcripts, data exports and the report), get the format, the producer and the delivery date in writing. Also state which outputs you will produce yourself. Ask what happens if recruitment falls short, covering replacements and who pays. A quote that says "qualitative study, four groups" next to one figure cannot be compared. Send it back and ask for line items.

Score each shortlisted agency on six weighted criteria using evidence you have seen, apply a gate that removes agencies weak on recruitment or ownership, then pilot the finalists. The scorecard below is our own model, not an industry standard.

How do you score agencies fairly?

Score each criterion from 0 to 5: 0 means no answer, 1 means assurance only, 3 means a documented process, and 5 means evidence you have seen. Weighted points are score ÷ 5 × weight, added across the six rows.

Gate: a score below 3 on ownership or recruitment removes the agency, whatever its total.

Worked example with illustrative scores, not real agencies. For Agency A: 4/5 × 20 = 16; 5/5 × 25 = 25; 3/5 × 20 = 12; 4/5 × 15 = 12; 3/5 × 10 = 6; 4/5 × 10 = 8; total 79.

A and B tie at 79, but B fails on recruitment, which the total hides. A and C are two points apart. At that margin the scorecard cannot choose, so a pilot decides.

What does a good pilot test?

A pilot is a small slice of the real study, run before you commit to the whole. Write the pass marks down before it starts.

• Recruits match the screener. Pass: every pilot participant meets every screener criterion. One mismatch means finding out why before you sign.
• Identity checks ran. Pass: the agency shows what was checked for each participant.
• Moderation. Pass: you watch or read one full session and the moderator probes without leading.
• Language. Pass: a bilingual colleague compares a transcript page with its translation and the meaning holds.
• No-shows. Pass: replacements come from the same quota cell with identical screening, and you are told.
• Deliverables. Pass: they arrive in the agreed format on the agreed date.

Ask whether the pilot is free, paid or credited against the full study. In our view, an agency that refuses any small pilot is telling you how it handles risk.

The questions that matter most on the first call are who recruits, who moderates, how identity is checked and where recordings are kept: an agency that cannot name people or give a retention period fails. Use these eight as a call script.

1. Who recruits and who moderates, by name? Fail: "our network", or names withheld until after award.
2. Which steps will you subcontract, and to whom? Fail: no answer, or an answer that changes between calls.
3. How is identity checked before invitation? Fail: "we trust our database."
4. What happens when you cannot fill a quota? Fail: quotas are loosened without telling you.
5. Which language will each session run in, and who transcribes? Fail: a language count with no named moderators.
6. Where do recordings go, and when are they deleted? Fail: no retention period.
7. What exactly do I receive, in which format, by which date? Fail: "a report" and nothing else.
8. Can we run a pilot first, and what will we both count as a pass? Fail: no pilot, or pass marks left vague.

Pressure to sign before you have seen a screener is also a red flag.

We are a recruitment partner that also supplies moderators and fieldwork support. We are not the only option, and the questions above apply to us too.

We recruit participants by city for in-depth interviews, focus groups, digital diary studies, ethnography, usability and app testing, online bulletin boards, in-home use tests, shop-alongs and opt-in digital behavioural tracking, and we supply moderators where a study needs them. Sessions can run in languages other than English. We run identity checks on participants before invitation and can target by household and demographic profile.

Depending on the project, studies are conducted by us, by research partners or by other research organisations. Deliverables, file formats and responsibilities are agreed before the study begins. Depending on scope, deliverables can include recruitment, fieldwork materials, recordings, transcripts, structured data exports and research reports. We produce what the engagement includes, and you produce outputs from research you manage yourself.

Choosing a qualitative research agency in India comes down to evidence you can see before you pay. Find out who recruits, who moderates and who owns each deliverable. Check the screener, the identity checks and the moderator’s language fit. Ask for the consent form, the retention period and a line-item quote that matches your brief. Then score two or three agencies on the same weighted criteria, apply the gate, and let a small pilot settle any close call. Write the pass marks down before the pilot starts. The scorecard is our own model, so adjust the weights to suit your study.

The same tests apply to us. If you would like to see how we answer them, email your brief to support@qualfacto.io and ask about the free pilot. We review the brief for feasibility and then send a tailored proposal and quote.