guide 4 min read
How to run a concierge test for a product idea
Use a concierge test to deliver a proposed product outcome manually, observe the real workflow, and distinguish useful learning from an unrepeatable service promise.
THE SHORT ANSWER
What you need to know
A concierge test delivers the proposed outcome through clear manual support before a complete product exists. Choose a narrow workflow, explain the support, and observe the inputs, decisions, and exceptions required to deliver it. Use the test to understand value and operational demands, then decide which parts could become product capabilities and which should remain human work.
Key takeaways
- Choose a narrow outcome that can be delivered honestly.
- Make manual support and participant expectations explicit.
- Log exceptions and invisible work during delivery.
- Separate product capability from service dependence.
Define a deliverable outcome and boundaries
A concierge test needs an outcome the participant can recognize, not an open-ended promise to solve every problem. Suppose a team wants to help community organizers prepare volunteer schedules. The test could produce one usable schedule from an agreed set of availability inputs. State what the team will prepare, what the organizer must supply, and what changes are outside the session. This boundary makes the test understandable and gives the team a consistent unit of work to study. It also prevents helpful improvisation from obscuring whether the proposed product actually supports a repeatable situation.
Observe the work instead of hiding it
Manual delivery can create a good experience while concealing important product difficulties. Record how information arrives, which details need clarification, and where the team makes judgment calls. The volunteer schedule may require several conversations because people express availability differently. That is not a minor setup detail if a future product depends on structured inputs. Keep a delivery log with steps, questions, exceptions, and revisions. Ask the organizer to explain what they considered useful and what they still had to do afterward. These observations show the whole outcome, including work outside the apparent product boundary.
Investigate value under realistic conditions
The participant’s reaction should be connected to a real task. Did the schedule help them coordinate the actual event, and what happened when a volunteer changed availability? Ask for the resulting workflow, not just a rating of the session. Explain when the support team is available so constant founder attention does not accidentally become the core value proposition. If participants rely on reassurance or judgment, that can be a valuable service finding. It simply changes the question from how to automate a schedule to which human and product capabilities the offer should combine.
Decide what deserves implementation
Review the delivery steps and classify them by the role they play: repeatable transformation, necessary judgment, participant communication, or unusual exception. A structured availability form might remove repeated clarification, while complex role assignment may remain a human decision. Test the next product component against that specific burden. The checklist is to define the outcome, state support clearly, collect realistic inputs, log manual work, follow the real result, and choose the next capability. Do not automate a step merely because it was tedious; first check whether it creates useful value or compensates for a poorly defined workflow.
After a delivery session, separate participant work from support-team work in the log. For volunteer scheduling, list which clarifications the organizer could reasonably provide and which decisions required the team’s judgment. Use the distinction to design the next test. A structured input step may reduce avoidable back-and-forth, while a judgment-heavy assignment step may need guidance rather than automation. Both are useful findings about the offer.
Manual delivery often exposes questions about inputs and workarounds. Use a focused, inexpensive research round to investigate those questions before automating the process you just observed.Research your first product without a big budget
Common questions
Is concierge testing the same as pretending a product exists?
No. Participants should understand that the outcome is being delivered through manual support. Clear expectations preserve trust and make the source of value easier to investigate.
What if participants mainly value personal attention?
Treat that as evidence about the offer. Investigate whether guidance is essential, whether the audience wants a service, or whether a product can support the judgment more effectively.
When should the team stop manual delivery?
Stop when the test has answered its planned question or when further sessions repeat the same learning. Continue only with a new question, narrower boundary, or component worth testing.
References & further reading
These primary references support the background. The practical frameworks and illustrative examples are written for this guide.
Prepared with AI assistance. Examples are illustrative; no client results are implied. Read our editorial approach.