Skip to main content

Tavism
All startup casesConcierge MVP

Concierge MVP & Manual Scaling

Food on the Table — Concierge MVP

Learn the customer's exact problem by doing the work yourself—before you write software that scales the wrong thing.

Case overview

The challenge

Manuel Rosso wanted a meal-planning service that matched grocery store sales with what families actually liked to cook. Building databases and recommendation engines first would hide whether anyone wanted the help.

The technique

A Concierge MVP with no software

Rosso visited one customer each week. He planned their meals, bought the groceries by hand, and delivered the plan in person. There was no automated software—only a founder doing the service so he could see the real workflow.

Scaling gradually

From one family to hundreds

He grew from one customer to ten, then to hundreds—still mostly manual—before writing automated database scripts. Each step confirmed the problem was real and showed which parts of the work deserved code.

1100s

Customers served before automation

Core principle

Do things that don't scale

Stay manual until you understand the customer's exact problem. Automation comes after the learning—not before you know what to encode.

For founders

Key takeaways

How to validate a service idea before you fund the product build.

  1. Deliver the service yourself until the customer's pain is obvious.
  2. A Concierge MVP skips software so learning stays fast and cheap.
  3. Scale headcount and process before you scale code.
  4. Automate only the steps you have already proven by hand.

Want a Concierge MVP before you build?

Book 30 minutes. Bring the service hypothesis you want to test with one real customer.

Design Your Concierge Experiment