Skip to content
Log in
← Statistics glossary

concierge MVP

Also called: concierge test, manual MVP

A first version where the founders deliver the result by hand, with spreadsheets, calls and messages, behind a simple front end. It tests whether customers value and pay for the outcome before anything is automated.

A concierge MVP is a first version of a product in which the founders deliver the result by hand, as a personal service, while the customer sees (at most) a simple front end. Spreadsheets, phone calls, messages and manual data entry stand in for the software that will one day do the job.

The point is to learn and to earn before building. Doing the work by hand shows exactly which steps customers value, which ones they skip and where the real effort lies, and the customer pays for the outcome from day one. It also avoids automating a process nobody has yet proven they need. The name comes from the idea of a hotel concierge: a person who takes care of things for you.

A concierge MVP doesn't scale, and it isn't meant to. Set a rule in advance for when to automate each manual step, for example once it takes more than a few hours a week, and keep a log of how long each task takes. The log becomes the specification for the first thing to build.

Example

At launch Nina types each GarageDesk shop's client list into the system herself; it takes about 3 hours per shop. She also counts by hand, once a month, how many cars came back after reminders, for Dennis's report.

In April 9 new shops join: 9 × 3 = 27 hours of typing in one month, far above the founders' rule to automate a manual task once it takes more than about 4 hours a week. So in May Nina builds a self-service import from a spreadsheet file, and she knows exactly which columns shops actually have, because she has typed in all of them.

Common mistakes

  • Hiding the manual work forever. Set the automation rule in advance, or the concierge becomes the bottleneck.
  • Doing it for free. A concierge MVP should still charge; otherwise you learn nothing about willingness to pay.
  • Not measuring the time. Without a log you can't tell which manual step to automate first.
  • Promising what can't scale. Don't sell a personal service you can't keep once there are fifty customers.