MethodsArticlesCompareAbout
MethodsArticlesCompareAbout
MethodsCo-Discovery Testing
TestingDesign & PrototypingQualitative ResearchIntermediate

Co-Discovery Testing

Uncover usability issues and mental models by observing paired users collaborating through tasks naturally.

Co-Discovery Testing pairs two users to explore a product together, revealing usability issues through their natural conversation and collaboration.

Share
Duration2 hours or more.
MaterialsAppropriate space, prototype, notebook, video camera.
People1 researcher, 6 or more participants.
InvolvementDirect User Involvement

Co-Discovery Testing is a usability evaluation method where two participants explore and interact with a product together while researchers observe. Also known as constructive interaction, the method leverages the natural conversation between paired users to surface usability issues, mental models, and workarounds more organically than traditional one-on-one think-aloud sessions. UX researchers, product teams, and usability specialists use this technique when they want richer verbal protocols without the artificiality of a single participant narrating their thoughts to a researcher. The paired format is especially valuable when testing products designed for collaborative use, when working with younger users who may feel intimidated by adult researchers, or when individual think-aloud sessions produce awkward silences. By listening to how participants explain features to each other, debate approaches, and troubleshoot problems together, researchers gain insights into how users naturally reason about and communicate product interactions. The method typically involves 4 to 6 pairs completing realistic task scenarios while a moderator observes, followed by a debriefing discussion that captures additional reflections and suggestions for improvement.

WHEN TO USE
  • When individual think-aloud protocols feel unnatural and produce limited verbal data from participants
  • When testing collaborative tools or products designed for shared use between two or more people
  • When working with participant groups who may feel intimidated in one-on-one researcher settings
  • When you want to understand how users explain and teach product features to each other naturally
  • When basic usability issues are resolved and you need deeper insights into mental models and workflows
WHEN NOT TO USE
  • ×When you need to measure individual task performance metrics like time-on-task or error rates precisely
  • ×When testing highly personal or sensitive products where social desirability bias between peers is a concern
  • ×When you have very limited participant availability since the method requires twice the recruitment effort
  • ×When early-stage prototypes have many basic usability issues that individual testing would catch more efficiently
HOW TO RUN

Step-by-Step Process

01

Define Objectives

Start by identifying the primary goals of your Co-Discovery Testing, including specific functionalities or aspects of the product that you want to assess. Be sure to clearly communicate these objectives to your team.

02

Select Participants

Choose a diverse group of participants who fit the target user demographics for your product. Aim for 4–6 pairs (8–12 people total), including a mix of novices and experts, or users with different background experiences.

03

Create Test Scenarios

Develop a set of realistic scenarios or tasks that individuals will need to accomplish using your product. These scenarios should cover key functionalities and user needs, aligning with your objectives. Be descriptive and detailed, providing context for the testers.

04

Introduce the Discovery Process

Explain to the participants how the Co-Discovery Testing process will work, detailing their roles, the dynamic of working in pairs, and the expected sharing of ideas and insights. Encourage them to ask questions and clarify any doubts.

05

Pair Participants

Pair participants together, considering the balance of knowledge and skills among them. Encourage productive conversation and collaboration during the testing process, ensuring that they are comfortable openly exchanging opinions and suggestions.

06

Conduct the Testing Sessions

Have participants work through your test scenarios as a pair, completing tasks while concurrently discussing their impressions, difficulties, and ideas for improvement. Encourage them to share their thoughts and feedback out loud throughout the session.

07

Observe and Record

During the testing sessions, have a moderator observe each pair of participants, noting their interactions with the product and with each other. Capture both verbal and non-verbal feedback, as well as any struggles, successes, or insights that emerge.

08

Conduct Debriefing

After the testing sessions, gather the participants to discuss their experiences and share any notable findings. This group discussion allows for additional insights and potential improvements to be identified and shared among the group.

09

Analyze Results

Review the data collected during the sessions, including observations, participant discussions, and debriefing notes. Analyze the findings to identify patterns, themes, and areas for improvement in the user experience.

10

Report Findings and Recommendations

Create a comprehensive report that outlines the key findings from the Co-Discovery Testing sessions, as well as actionable recommendations for improving the overall user experience. Share the report with your team and stakeholders for implementation of improvements and future testing.

EXPECTED OUTCOME

What to Expect

After running Co-Discovery Testing sessions, your team will have rich qualitative data about how users naturally interact with and reason about your product. The paired conversations will reveal usability issues, confusing interface elements, and unexpected workarounds that participants discover together. You will gain insight into how users explain product features in their own language, which highlights both strengths and problem areas in your design. The analysis will produce a prioritized list of usability findings, documented user mental models, and actionable design recommendations. Teams typically find that co-discovery produces more candid and detailed feedback than individual testing, especially for collaborative tools and products targeting younger demographics.

PRO TIPS

Expert Advice

This method works well when a power dynamic between researcher and participant might inhibit natural responses.

Review video recordings with participants afterward to better interpret their communication and decision-making.

Consider group testing as an alternative that leverages interaction dynamics even more -- useful as part of focus groups.

Pair participants with similar expertise levels to ensure balanced conversation and mutual discovery.

Brief pairs on encouraging each other to think aloud rather than working silently side by side.

Position yourself where you can observe without interrupting the natural pair dynamic or conversation flow.

Listen for moments when one participant explains something to the other -- these reveal critical assumptions.

Use pairs who know each other for more natural conversation, or strangers for unbiased fresh perspectives.

COMMON MISTAKES

Pitfalls to Avoid

Mismatched expertise levels

Pairing a novice with an expert creates a teaching dynamic rather than collaborative discovery. Match participants by experience level to ensure balanced contribution and genuine problem-solving together.

Intervening too often

Moderator interruptions break the natural conversation flow between participants. Stay silent unless participants are completely stuck, and save questions for the debriefing session.

Ignoring social dynamics

One participant may dominate while the other becomes passive. Watch for this pattern and gently encourage both participants to share their thoughts during the session introduction.

Not recording conversations

Relying on notes alone misses critical micro-interactions between participants. Always video record sessions to capture both verbal exchanges and non-verbal cues for thorough analysis.

Skipping the debrief

The debriefing discussion often reveals insights that did not surface during tasks. Always allocate time for participants to reflect on their experience and discuss what they noticed.

DELIVERABLES

What You'll Produce

Test Plan

Plan outlining objectives, scenarios, participant profiles, and data collection methods.

Test Script

Step-by-step task guide ensuring consistency across all participant pairs.

Participant Consent Forms

Signed forms covering study purpose, confidentiality, and participant rights.

Data Collection Tools

Video/audio recorders, notepads, and questionnaires for capturing interactions.

Data Analysis Framework

Approach for analyzing data using affinity diagrams or behavioral coding.

Findings and Recommendations Report

Report detailing findings, patterns, and actionable UX recommendations.

Test Results Presentation

Visual presentation of key findings tailored for stakeholder audiences.

Screener Survey

Pre-screening questions to recruit suitable participants for testing pairs.

FAQ

Frequently Asked Questions

METHOD DETAILS
Goal
Design & Prototyping
Sub-category
Usability testing
Tags
co-discovery testingusability testingpair testingcollaborative testingconstructive interactionthink alouduser testingqualitative researchusability evaluationuser behavior
Related Topics
Usability TestingThink-Aloud ProtocolCollaborative DesignUser-Centered DesignQualitative Research MethodsParticipatory Design
HISTORY

Co-Discovery Testing, also called constructive interaction, was introduced by researchers Penny Miyake and Don Norman in the 1980s as an alternative to the traditional think-aloud protocol developed by Clayton Lewis. The method was born from the observation that having a single participant verbalize their thoughts while performing tasks often felt unnatural and could interfere with task performance. By pairing two users, the conversation that naturally occurs between them replaces the artificial think-aloud monologue. The approach gained recognition in the HCI community through studies demonstrating that paired users produced richer verbal protocols and more natural interactions than individual think-aloud participants. Over the decades, the method has been particularly adopted in educational technology research, children's interface design, and collaborative software evaluation, where its strengths in capturing social interaction dynamics are most valuable.

SUITABLE FOR
  • Testing physical and digital artifacts through collaborative exploration
  • Evaluating products designed for younger users who may be reserved with adult researchers
  • Later iterations of testing when basic usability issues have been resolved
  • Understanding how users naturally explain features and troubleshoot together
  • Discovering social and collaborative aspects of product usage patterns
  • Testing products intended for pair or team use in their natural context
  • Generating richer verbal protocols through natural conversation between participants
  • Reducing participant anxiety by providing peer support during testing sessions
RESOURCES
  • Usability FirstCo-discovery method or constructive interaction is a user-testing technique where users work collaboratively to solve problems on a computer
  • Design Method Toolkit Co-discovery
  • Co-discovery Learning
  • What is a co-discovery usability session? - UX Insights Weekly Video TutorialCo-discovery usability sessions use more than one participant and rely on listening to how these participants communicate in order to work out what needs to be fixed in the user experience. Learn what type of product and users these sessions work best with, and how to set up your own co-discovery usability test.
  • Discovery phase: UX activities & research methods to considerUX Activities and Research Methods to apply in Discovery to understand the users, investigate opportunities, and facilitate a shared vision amongst stakeholders.
RELATED METHODS
  • Card Sorting
  • Design Sprint
  • Design Studio Method

93 methods. Step-by-step guides. No signup required.

ExploreAll MethodsArticlesCompare
PopularUser TestingCard SortingA/B TestingDesign Sprint
ResourcesAboutArticles & GuidesQuiz

2026 UXAtlas. 100% free. No signup required.

93 methods. Step-by-step guides. No signup required.

ExploreAll MethodsArticlesCompare
PopularUser TestingCard SortingA/B TestingDesign Sprint

2026 UXAtlas. 100% free. No signup required.