Singapore founder discovery: individual interviews or a joint shared-task account?

The problem

Your prospective users perform different parts of the same task, and interviewing one spokesperson or combining their joint account may leave you unclear about whose experience and how many actual episodes your notes establish.

Choose the interview unit from the shared task

When two people actually use a service together, decide whether you need an individual account of each person’s part or a joint account of their interaction. Two colleagues in the same room do not automatically constitute one shared task. Name the activity, the particular episode you want to understand and what each person does within it. Use the founder guide to ask about a recent task before pitching, and the user/buyer/approver guide to separate task experience from adoption authority. Plan the conversation around how the work itself crosses between users. A joint conversation can be relevant to that question without becoming two independent end-to-end accounts or a collective buying decision.

Use the source’s method distinction at its stated scope

The UK GOV.UK Service Manual’s in-depth-interview guidance, published on February 21, 2017, says interviews normally involve one user and can involve pairs or small groups when people use a service together, such as team members sharing tasks. It also calls for current or likely users, research questions, a discussion guide and participant consent. Use that as attributed method context, not a Singapore requirement or a rule that joint sessions always work better. An individual account may help examine one person’s experience; a joint session may help discuss the shared activity. Explain the question each arrangement is intended to address, and establish its appropriate recruitment and permissions through your actual research process.

• GOV.UK Service Manual: using in-depth interviews ↗ · Source checked 21 October 2026

Plan the handoff before choosing a spokesperson

Build a short record with six entries: shared task and episode; actors and each person’s part; action or handoff you need to understand; whose account or observation is currently available; individual or joint arrangement proposed and why; and missing perspective or condition. Use this suggested record to keep each person’s part and the question for the proposed conversation visible. A person who receives an item may describe receiving it while knowing little about the next person’s update. A manager who approves buying a system may not perform either step. Keep those differences visible. Do not nominate the most senior or most talkative attendee to supply every user’s experience. A joint invitation is still only a proposed session until the actual participants and research arrangement are established.

Reconstruct one invented key-return episode

Every person, key, time and account here is fictional. A founder wants to understand a shared key-loan return task. In the supplied practice material, desk worker Ren reports receiving key K7 at 09:00. Inventory worker Mei reports updating that key’s availability at 09:15. Both accounts concern the same return episode, but each describes a different part. An individual conversation with Ren could examine the receiving step; it leaves Mei’s update account absent unless obtained separately. A proposed joint conversation could ask the two users to describe the handoff between their parts. That choice follows the shared work rather than their job titles. The material contains no direct observation of the original return and no explanation for the interval.

Retain each contribution inside the joint account

Continue the invented case with a supplied joint discussion in which Ren describes handing K7 to the return tray and Mei describes checking that tray before updating the record. Keep those statements attached to their respective speakers and the same episode. They add reported steps; they do not turn the earlier return into something the founder directly watched. The record now has two users’ contributions to one reconstructed episode. Reporting two independently observed returns, or saying both people completed every step, would change the supplied evidence. The fifteen-minute interval still has no established cause. Do not label the key lost or the workflow broken from those times alone. If accounts conflict, preserve the difference and the missing check instead of manufacturing a unanimous story.

Take the session-choice question to an appropriate learning format

For a founder-learning event, ask whether the actual exercise helps plan customer research, examines a permitted participant case or merely rehearses a pitch. You might bring the clearly fictional two-role record and ask, “Which missing part calls for an individual account, and what could a joint reconstruction examine without changing what our current notes establish?” That is a methods question, not a request for attendees to act as your actual customers. Confirm the preparation, accepted material and discussion route. If the event uses only supplied exercises, retain that origin. A facilitator’s recommendation can improve the proposed research arrangement while providing no missing user statement, consent or observed task. Keep a workshop place separate from recruitment for later real research.

Choose the next permitted account without expanding the finding

Finish with three lines: the shared-task question; the individual or joint arrangement you propose; and the evidence that arrangement would still need to supply. In the fictional case, a joint discussion may help explore the handoff, while the cause of the interval and any adoption decision remain open. Use the existing founder role guide if the missing question concerns purchase approval, and the UX sample or journey guides for those separate claim limits. Search Boardzip’s free startup search for an actual discovery-learning or research-planning candidate, then confirm its participant activity and access. If none offers the required discussion, retain the plan for a suitable opportunity. A search result supplies neither a research participant nor permission to interview a group.