The important difference is intervention
In a moderated session, the researcher can clarify and ask a relevant follow-up while the participant is present. In an unmoderated session, the task and software need to carry more of that instruction. NN/g describes this distinction and explains why ambiguous tasks can be particularly expensive in unattended studies. [S19]
For a complex unfinished flow, clarification may be necessary to continue. For a straightforward task on a functional design, repeated intervention may instead change what you are trying to observe. Write down which kind of interaction your research question needs, rather than choosing the format solely because one price is lower.
Published live-interview routes in this edition
Lookback documents live modes alongside Tasks and SelfTest. Useberry now documents moderated interviews on paid plans. Maze documents Interview Studies and moderated recruitment as Enterprise features. These are product-scope facts, not comparative evidence that any one moderator or platform produces better research. [S34] [S06] [S15]
Useberry’s cited interview recruitment instructions document an own-participant share-link workflow and describe panel integration as forthcoming. Therefore, the picker does not promise that its unmoderated participant pool can currently fill a live interview. A platform offering recruitment somewhere does not mean the same route is available for every method. [S07]
Unmoderated does not mean numbers only
Lookback’s documentation describes qualitative unmoderated recordings and AI follow-up prompts. Userbrain presents its task-based recording workflow. A transcript, recording, path metric and AI summary are different outputs; choose the output you intend to inspect, not just the largest feature list. [S35] [S36]
Keep synthetic participants separate from real people taking a study. AI assistance that helps organise or prompt a real session is not evidence from an additional human participant. Our catalogue does not turn generated personas into a lower-priced human research panel.
Compare the complete workflow
For a live session, budget for coordinating availability, attendance, preparation and review. For an unattended session, budget for task design, participant instructions, recordings and analysis. The point is not to assert that one always takes longer; it is to keep work that moved outside the session from disappearing from the estimate.
Write the deliverable in advance. It may be a set of clips explaining a failure, a count of task outcomes, a revised interview guide, or a comparison of two versions. Then verify that the documented export and collaboration route on the chosen plan supports that deliverable.
Use the smallest sufficient shortlist
Choose “live interview,” “recorded task,” “structured prototype” or “first click” in the tool picker. Its result is an explained, documentation-based shortlist. Unknown plan behavior does not become a fabricated “no,” and no score claims to predict research quality.
For a more concrete live-versus-recorded purchase, read Lookback versus Userbrain. For the cost components that remain after choosing the method, use the budget guide.