TL;DR
Choose UX research agencies by scoring participant quality, method fit, analysis depth, actionable outputs, and commercial terms. Use the same brief for every finalist, require evidence for each score, and reject any agency that scores below 3 out of 5 for recruitment, method fit, or evidence traceability.
Score each shortlisted UX research agency across 5 criteria before you choose one: participant quality, method fit, analysis, actionable outputs, and commercial terms. Appoint the highest-scoring agency only if it can show how its recommendations will trace back to the right users and observed evidence.
Use One Brief and One Scorecard
A shared scorecard stops the loudest sales call from deciding your research partner. Send every agency the same one-page brief, then score the proposal and pitch against the same requirements.
Use this matrix during shortlisting and again after finalist calls.
| Criterion | Weight | Score 0 | Score 5 |
|---|---|---|---|
| Participant quality | 30 | No screener or source | Screener, quotas, verification, replacements |
| Method fit | 25 | Method list only | Decision, method, limits clearly linked |
| Analysis depth | 20 | Findings without evidence | Findings traceable to participant evidence |
| Actionable outputs | 15 | Slide deck only | Priorities, owners, readout, next actions |
| Commercial terms | 10 | Unclear scope | Itemised scope, roles, data terms |
Score each row from 0 to 5. Calculate the weighted result as (score ÷ 5) × weight.
For example, an agency scoring 4, 4, 3, 4, and 4 receives:
24 + 20 + 12 + 12 + 8 = 76 out of 100
Our rule: reject any agency that scores below 3 for participant quality, method fit, or analysis depth. A polished presentation cannot rescue the wrong participants or a method that cannot answer your product question.
Set the Decision and User Definition
Write the decision before you contact agencies. “We need user research” is too broad to evaluate a proposal.
Use this brief:
- Decision: What will the team choose after the study?
- Workflow: Which task, feature, journey, or concept is involved?
- Users: What must participants have done recently?
- Segments: Which roles, markets, devices, or accessibility needs matter?
- Constraints: What release date, prototype maturity, legal review, or language requirement applies?
- Useful output: What should product, design, or leadership do differently?
A strong brief says, “Choose which onboarding flow to build for finance managers who approved a payment in the past 30 days.” A weak brief says, “Understand our B2B users.”
Prioritise recent behaviours over job titles. A person with the right title may never perform the workflow you need to study. NN/g’s recruiting guidance also recommends defining distinct user groups before recruiting.
If participant access is your main risk, use this India recruitment cost worksheet to separate recruitment, incentives, and research-delivery costs.
Test the Method, Not the Method List
An agency should recommend a method because it answers your decision, not because it is the agency’s favourite service.
Use interviews when you need to understand motivations, language, workarounds, or unmet needs. NN/g describes user interviews as a discovery method that helps teams learn about users before deciding what to build.
Use usability testing when you need to watch people complete tasks in a prototype, website, or app. Ask the agency to define the tasks, success criteria, prototype state, and what a failed task would mean for the roadmap. For qualitative usability testing, NN/g’s starting guidance is 5 participants, with exceptions for distinct user groups and other study types.
Use diary or field-based research when behaviour unfolds over days or depends on the user’s environment. NN/g’s method guide includes diary studies and field studies alongside interviews and usability testing.
Ask each agency these 3 questions:
- Which decision will this method answer?
- What will this method fail to answer?
- Why is this method better than the closest alternative?
A useful answer names a trade-off. For example, an agency might recommend interviews to understand approval-workflow workarounds, then usability testing to check whether a new flow solves them.
For live task observation, see our guide to when moderated usability testing pays off. For longer-term behaviour, use our diary study planning guide.
Inspect Recruitment and Data Practices
Recruitment quality decides whether your findings describe your users or strangers who fit a loose demographic. Ask to see the screener before fieldwork begins.
The screener should test relevant behaviour, tools used, frequency, decision authority, and exclusions. Ask where participants come from, how the agency checks eligibility, and what happens when a participant does not attend.
Ask for a replacement policy with 3 specifics:
- The point at which an unsuitable participant is replaced.
- The time allowed to fill a cancelled session.
- Who pays when a screener fails.
Also ask how recordings, contact details, consent records, and raw notes will be handled. The ICC/ESOMAR research code requires clear communication about personal-data collection and use, plus protection against unauthorised access.
Your proposal should name who controls participant data, where recordings are stored, who can access them, and when materials are deleted. Those details matter most for sensitive workflows, regulated industries, and research involving employees or customers.
Read the Proposal for Evidence
A useful proposal is a working research plan. It should make it easy for your team to see what it is buying and what decisions it will support.
Look for these 8 parts:
- The product decision and research questions.
- Participant definition, screener, quotas, and recruitment source.
- Method rationale and session format.
- Named moderator, analyst, and project lead.
- Analysis process and how findings connect to evidence.
- Deliverables, including readout and decision workshop.
- Timeline, dependencies, and scope-change process.
- Itemised costs for design, recruitment, incentives, fieldwork, analysis, readout, travel, and change requests.
Request an anonymised deliverable from a comparable project. Look for a finding, the participant evidence behind it, the product implication, and the priority assigned to it.
A slide deck full of quotes is not enough. Your product team needs a record that answers: what happened, how often it mattered in the study, why it happened, and what the team should change.
Ask These Questions Before Appointing an Agency
Use the finalist call to test whether the people selling the work can reason through your study.
- Which product decision will change after this research?
- How will you prove each participant fits our target user definition?
- Who writes the screener and who approves it?
- Who moderates the sessions and who performs the analysis?
- Show us one anonymised finding and the evidence behind it.
- How will you distinguish a one-off comment from a prioritised finding?
- What will product, design, and leadership receive after the study?
- What is included in the fee, and what triggers extra cost?
- How will you obtain consent and handle recordings or identifiable data?
- What will you do if the research invalidates our preferred solution?
The final question is revealing. A capable research partner explains how the study can challenge the brief, not only how it can validate a proposed answer.
Where Qualfacto Fits
At Qualfacto, we manually review and match participants for focus groups, in-depth interviews, and UX tests, and our research formats also include digital diary studies and at-home ethnography. If recruiting Indian participants is the critical risk in your study, start with our qualitative research panel.
FAQs
How Many UX Research Agencies Should I Shortlist?
Shortlist 3 agencies. Three proposals give your team a meaningful comparison without creating a procurement process that delays the research decision.
Use the same brief and scorecard for all 3. Invite only agencies that can work with your target users, required market, and study timeline.
Can I Hire One Partner for Recruitment and Another for Research?
Yes. Separate recruitment and research delivery when one provider has stronger access to your audience and another has stronger study-design capability.
Make the handoff explicit. Agree on the screener owner, consent language, participant contact rules, scheduling process, and data access before recruitment starts.
Should I Run a Paid Pilot Before a Larger Programme?
Yes, when the audience is difficult to recruit, the research is sensitive, or the agency’s working style is unknown. A pilot should test one decision, one participant segment, and one deliverable format.
Score the pilot with the same matrix. The pilot should show how the agency recruits, moderates, analyses, and turns evidence into action.
What Should a UX Research Agency Contract Cover?
The contract should cover scope, participant recruitment, incentives, deliverables, timing, change requests, ownership, recordings, consent, confidentiality, and data retention. The proposal should identify the people responsible for moderation and analysis.
Treat those clauses as part of research quality. A vague contract creates vague responsibilities when recruitment or scope changes.
