The Concierge MVP, Explained
A concierge MVP is the version of your product where you, the founder, are the product. Instead of building software and hoping people use it, you deliver the outcome by hand to a small number of real customers, personally and openly, and watch what they actually need. It is slower per customer than any automated approach, and that is exactly the point: doing the work by hand is how you learn what the software will have to do, what people will pay, and whether the whole idea holds up under real use. This guide covers what a concierge MVP is, how it differs from the other early-stage techniques, how to run one step by step, and how to know when it is time to write code.
What a concierge MVP is
The name comes from a hotel concierge: a real person who takes your request and handles it for you. In a concierge MVP you do the same for your customers. You take their input, you do the work manually with whatever tools you already have (a spreadsheet, email, a phone call, a shared document), and you hand back the finished result as if the product had produced it. The customer knows a person is involved. There is no pretense of automation.
Say your idea is an app that builds a weekly meal plan from someone's fridge and goals. A concierge version skips the app entirely. The customer sends you a photo of their fridge and their goals; you write the plan by hand in a document and send it back. If they come back next week, and especially if they pay, you have learned something no survey could tell you: that the outcome is worth having, delivered this way, to this person.
Concierge vs wizard of oz vs landing page
These three techniques are often confused, but they answer different questions and reveal themselves to the customer very differently. Here is how they compare.
| Technique | What the customer sees | Question it answers | Best when |
|---|---|---|---|
| Landing page | A page describing a product that does not exist yet | Will anyone want this? | Your risk is demand |
| Concierge | A person openly doing the work for them | Is the value real, and what shape is it? | Your risk is whether the outcome matters and what to build |
| Wizard of oz | What looks like finished software | Will people use it as a product, unassisted? | You need to test the real interface and behavior |
If you want the full detail on the third one, we cover it in the wizard of oz MVP, explained. Many products move through all three in order: a page to test the pull, a concierge phase to learn the work, then a wizard of oz phase to test the interface before anything is fully automated.
When to run a concierge MVP
Choose this technique when your biggest unknowns are on the delivery side rather than the demand side. It shines in a few situations.
- The value depends on judgment or quality that is hard to specify up front. Doing it by hand is how you discover the rules you would otherwise have to guess at.
- You are not yet sure what the customer actually wants delivered, only that they have a problem.
- Building even a rough version would be expensive, and you want to be sure before committing.
- The service is high-value and low-volume, so serving a few people by hand is genuinely feasible.
It is a poor fit when the product only works at scale (a social network is no fun with one hand-served member), when the value is speed no human can match, or when customers would never accept a human in the loop for privacy reasons.
How to run a concierge MVP, step by step
The discipline is to stay manual on purpose and to treat every interaction as data. Here is the sequence we recommend.
- Define the one outcome. Write the single result the customer is paying for in one sentence, the same way a good MVP spec names one job. Everything you do by hand should produce that outcome.
- Find five to fifteen real customers. Not friends. People with the problem, ideally sourced the way you would source them at scale, so what you learn transfers.
- Charge something. Even a small or discounted price. Payment is the difference between "I would use that" and "I am using that."
- Deliver by hand, and keep the toolkit crude. Spreadsheets, email, documents, calls. Resist building anything. The absence of tooling is what keeps you close to the work.
- Write down every step you take. Each manual action you repeat is a future feature. Each question you have to ask is a future form field. Each judgment call is a future rule.
- Watch where it hurts. Note where you spend the most time, where customers get confused, and where they ask for something you did not plan.
- Ask for the renewal. Repeat use, and especially repeat payment, is the clearest signal that the value is real.
What a concierge MVP teaches that a landing page cannot
A landing page tells you people will click. A concierge MVP tells you what happens after they click, and that is where most product ideas quietly fall apart. Serving customers by hand surfaces the things you cannot imagine from a desk: the edge cases, the input people actually have rather than the input you assumed, the moment they lose trust, and the parts of the outcome they care about versus the parts you were proud of that nobody notices.
It also hands you a specification written in reality rather than in speculation. By the time you have served a dozen customers, you know the exact steps, the exact inputs, and the exact decisions the software must make, because you have made them yourself, repeatedly. That is a far stronger foundation for a build than any brainstorm, and it directly connects to sound feature prioritization, because you have felt which steps matter.
Pricing from day one
Founders often want to give the concierge service away to "get feedback," and it is almost always a mistake. Free customers are polite and forgiving; paying customers are honest and demanding, and demanding is what you need. Charging early does three things at once: it validates that the outcome has real economic value, it teaches you what the market will bear, and it filters out the well-wishers so your learning comes from genuine buyers. If a price makes people walk away, that is a finding, not a failure.
Knowing when to stop being the concierge
A concierge MVP is a phase, not a business model, because your own hours are the ceiling. You are ready to move to software when the manual work has become repetitive and predictable, demand is steady enough to justify the investment, and your time has clearly become the bottleneck to serving more people. When you know the exact steps well enough that you are basically following your own script, you are ready to automate.
From concierge to code
When the manual version is running smoothly and you know precisely what to automate, the build itself becomes low-risk, because the hard questions are already answered. This is the point where a fixed price is genuinely possible: the scope is bounded by everything you learned serving customers by hand. To get a market number for that build, run your feature list through the free MVP cost calculator, and see how the routes compare in how much an MVP costs in 2026.
At SquadPrime we build that first automated version at a fixed price from $4,000, staffed by senior professionals, with you owning 100 percent of the code. A concierge phase is the best possible preparation for it, because you arrive with a spec drawn from real customers rather than assumptions, and that is what makes a single confident quote safe to give.
Have a spec? Get one fixed price for it.
Bring your spec to a free 30-minute call. We will pressure-test the scope, flag anything that will inflate the build, and quote one fixed price (from $4,000). Live in 14 days or you don't pay, and you own all the code.
Get My Launch Date & Fixed PriceFAQ
What is the difference between a concierge MVP and a wizard of oz MVP?
In a concierge MVP the customer knows a person is serving them by hand; the manual work is visible and personal. In a wizard of oz MVP the customer believes they are using finished software, while people do the work invisibly behind the scenes. Concierge is honest and high-touch; wizard of oz simulates automation the user cannot see.
Do I charge customers during a concierge MVP?
Usually yes, if you can. Charging from day one is the strongest possible signal, because it proves people value the outcome enough to pay, not just to say they are interested. Even a small or discounted price filters real customers from polite well-wishers and teaches you what the market will bear.
How many customers do I need for a concierge MVP?
Fewer than you think. Because the method is high-touch, five to fifteen real customers is often enough to see clear patterns in what they need, where they get stuck, and what they will pay. The goal is depth of learning per customer, not a large sample, so you serve a small group extremely well.
When should I stop doing things manually and build software?
Stop when the manual work is repetitive, predictable, and clearly worth automating: you know the exact steps, demand is steady, and your own time has become the bottleneck. At that point you have a proven, detailed spec for the build, which is exactly what lets a team quote a fixed price with confidence.
All third-party names and trademarks belong to their respective owners and are used for identification only. Any market figures reflect our experience building MVPs and typical ranges as of September 21, 2026, and may change at any time. Spotted something outdated or inaccurate? Email talk@squadprime.com and we'll correct it promptly.
Related: The Wizard of Oz MVP, explained · The Landing Page MVP, explained