Skip to main content

Choosing Moderated vs Unmoderated Research in Lookback (Operational Differences)

Choosing Moderated vs Unmoderated Research in Lookback (Operational Differences)

Written by Henrik Mattsson
Updated over 2 months ago

Lookback supports both moderated and unmoderated research - but the difference between them is not just presence vs absence of a moderator.

They differ operationally in:

  • how participants are guided

  • how data quality is ensured

  • how sessions are run and observed

  • how researchers intervene (or don’t)

This article explains how to choose between them in practice, not in theory.


Moderated research in Lookback

Moderated research involves a researcher guiding the session in real time.

In Lookback, moderated research is conducted using:

  • LiveShare – continuous participant screen sharing

  • Interview – conversation-focused, limited screen sharing

When moderated research works best

Choose moderated research when you need to:

  • explore why users behave a certain way

  • probe unclear or unexpected behavior

  • adapt questions in real time

  • guide participants through complex flows

  • involve stakeholders live during sessions

Moderated research is especially effective when:

  • the product or prototype is complex

  • the research questions are exploratory

  • participant interpretation matters more than speed or scale


Unmoderated research in Lookback

Unmoderated research allows participants to complete sessions on their own time.

In Lookback, unmoderated research is conducted using:

  • Tasks – structured flows with step-by-step prompts

  • SelfTest – lighter-weight, instruction-driven sessions

When unmoderated research works best

Choose unmoderated research when you need to:

  • reach participants across time zones

  • collect data at scale

  • observe natural, uninterrupted behavior

  • reduce scheduling overhead

  • run diary or longitudinal studies

Unmoderated research is still qualitative in Lookback:

  • screen, audio, and often camera are recorded

  • participants are encouraged to think out loud

  • sessions stream live as they happen


Key operational differences

Researcher presence

  • Moderated: Researcher is present and guides the session

  • Unmoderated: Researcher is not present; instructions and tasks guide behavior


Data quality control

  • Moderated:

    • researcher can clarify misunderstandings

    • follow up immediately

    • adapt questions on the fly

  • Unmoderated:

    • relies on clear instructions and task design

    • can use AI moderation in Tasks to prompt clarification and elaboration


Participant experience

  • Moderated:

    • more conversational

    • participants can ask questions

    • pacing is researcher-driven

  • Unmoderated:

    • self-paced

    • participants follow instructions independently

    • pacing is participant-driven


Device and setup considerations

  • Moderated LiveShare on mobile:

    • requires the Participate app

    • enables full screen recording

  • Moderated Interview on mobile:

    • no app required

    • no screen sharing

  • Unmoderated on mobile:

    • requires the Participate app

    • supports screen, audio, and camera recording


AI-assisted research considerations

Lookback’s AI plays different roles depending on the method:

  • Moderated research:

    • AI assists with note-taking and surfacing moments of interest

    • researchers remain fully in control of the session

  • Unmoderated research (Tasks):

    • AI moderation can prompt follow-ups

    • helps ensure participants fully answer questions

    • reduces common unmoderated failure modes

AI supports researchers - it does not replace them.


Choosing the right method: practical guidance

Ask yourself:

  • Do I need to adapt questions in real time? → Moderated

  • Do I need scale or async flexibility? → Unmoderated

  • Is misunderstanding likely without clarification? → Moderated or AI-moderated Tasks

  • Am I running a diary or longitudinal study? → Unmoderated

Many teams use both approaches together within the same Project.

Did this answer your question?