The Manual MVP: Fake the Product Before You Build It
A manual MVP means you deliver the exact result your product promises by doing all the work by hand yourself, before you build any software, so you learn whether people actually want the outcome without spending a single weekend coding. The customer gets what they paid for. They just never see that “the app” is a 16-year-old with a spreadsheet and a phone.
MVP stands for “minimum viable product,” the smallest version of your idea a real person can use and judge. Most founders picture a stripped-down app. But the truly minimum version is often no app at all — it’s you, quietly delivering the service by hand and learning everything the finished product would need to do.
This is called a concierge MVP, because you act like a hotel concierge doing everything personally for each customer. It’s one of the highest-leverage moves a broke high-school founder can make, and this post shows you exactly how to run one.
What is a concierge MVP, really?
A concierge MVP is a test where you manually perform the service your product is supposed to automate, one customer at a time, for real money or a real commitment. You are the product. The “software” is you.
Say you want to build an app that generates weekly meal plans for busy students on a $40 grocery budget. The automated version takes months to build and might get everything wrong. The concierge version takes one afternoon: you find three students, ask about their week and their budget, and text each of them a hand-made meal plan every Sunday night. No app. No code. Just you, delivering the actual result.
Here’s the mindset shift. You’re not building a smaller product — you’re becoming the product temporarily so you can watch it work. Every question a customer asks, every plan they ignore, every “can you also do X?” is a spec for the real software. You’re getting paid to run your own research study.
Concierge MVP vs. wizard of oz vs. fake door
People mix these three up constantly, so let’s sort them out. They all fake a finished product, but they answer different questions.
| Test | What the customer sees | Who does the work | Question it answers |
|---|---|---|---|
| Fake door | A button or page for a product that doesn’t exist yet | Nobody — it just captures a click | Do people want this enough to reach for it? |
| Concierge MVP | A person openly helping them, no software claimed | You, by hand, and they know it | Can I actually deliver the result, and will people pay for it? |
| Wizard of oz | What looks like a working app or bot | You, by hand, hidden behind the scenes | Will people use the real experience if it existed? |
The fake door test measures demand before you’ve built anything. A concierge MVP goes further: you actually deliver the outcome, so you learn whether you can make people happy and whether they’ll pay. The wizard of oz version is the sneaky cousin — the customer thinks a computer is doing the work, but it’s secretly you typing in the back.
For a first-time founder, concierge is usually the best of the three. It’s honest, cheap, and teaches you more per customer than any survey ever will.
Why do this instead of just building the app?
Because building the app first is the single most expensive way to find out you were wrong. A concierge MVP buys you four things code can’t.
It kills the “would you use this?” trap. Ask people about a hypothetical app and they’re polite and useless — the reason your friends saying they love it is not validation. But when someone hands you $10 for a meal plan you’re about to make by hand, that’s not a compliment. That’s a purchase.
It teaches you the real product spec. You think you know what to build. You don’t, not yet. Maybe the meal plan is fine but customers really wanted a shopping list sorted by grocery aisle. You’d never guess that from a whiteboard. Now it’s the core feature.
It costs almost nothing. No developer, no subscriptions, no money you don’t have — just your time and a way to reach customers.
It gives you proof for later. Ten happy customers served by hand is real traction, and that story is exactly what to put on a traction slide when you have no revenue.
How to run a concierge MVP, step by step
You can start this week. Here’s the whole loop.
- Write down the exact outcome you’re selling. Not the app, the result: “A student gets a week of $40 meals, texted every Sunday.” If you can’t say it in one sentence, you’re not ready to test it.
- Find three to five real customers. Strangers, not just friends, because friends are too nice to teach you anything. Post in a school Discord, a subreddit, a group chat two degrees removed from you. If you’re stuck, start with where to find people to interview — the same places hold your first customers.
- Charge something, even if it’s small. Five dollars is fine. The point is to make them do something that costs them a little, so their “yes” is real. That’s the logic behind preselling before you build.
- Deliver the result by hand. Spreadsheet, texts, a Google Doc, a phone call — whatever gets them the outcome. Slow and manual is the point. You’re the engine right now.
- Watch everything and take notes. What did they ask for? What did they ignore? Where did they get confused? Every friction point is a feature the real product will need. This is collecting feedback that actually improves your product, done live.
- Do it again, better. Serve the next customer with what you learned. Repeat until you hit a clear “people love this and pay” or a clear “nobody cares.” Both answers are gold.
Notice there’s no coding step. That’s not a mistake. You don’t touch software until people are practically pulling the service out of your hands.
How do you know when it’s time to actually build?
You build automation when doing it by hand becomes the bottleneck, not before. The signal is simple: you have paying customers who keep coming back, and you literally cannot serve more of them by hand fast enough.
That’s when code earns its keep. Now you’re not guessing, because your manual work is the blueprint. You know the exact steps and the things people don’t care about, so you can build a lean no-code MVP that does only what your hand-work proved matters. You skip the six features nobody asked for.
If you never hit that bottleneck, that’s an answer too. If serving five customers by hand is easy because barely anyone wants it, you just saved yourself months of building the wrong thing. A fast no is a gift.
Watch for the false version of “ready,” though. Being tired of manual work is not the same as being blocked by demand. If you’re bored of texting meal plans but only three people ever bought one, more automation won’t fix that — that’s a demand problem, and no app solves a demand problem.
The mistakes that ruin a concierge test
A manual MVP only works if you don’t quietly cheat it. Watch for these:
- Serving only friends. They’ll say yes to be kind and teach you nothing. You need a few strangers who owe you nothing.
- Doing it all for free. Free hides the real question. If nobody will part with even $5, you don’t have a business yet, you have a favor.
- Automating too early. The urge to “just build a little tool” hits fast. Resist it. Every hour you code before serving real customers is an hour spent guessing.
- Not taking notes. The whole value is what you learn while delivering. Skip the notes and you throw away the point of the test.
- Testing with one person. One customer is an anecdote. Three to five gives you patterns you can trust.
Where this fits in building your company
The concierge MVP is a Build move, but it’s really a bridge. It sits right after you’ve validated that a problem exists and right before you build anything real. It’s the cheapest way to learn what to actually make while it still costs almost nothing to change direction.
This hand-first approach is baked into the Build sprint inside the batch0 program. Instead of rushing teen founders into an app they’ll regret, we push you to deliver your idea by hand to real people first, so the thing you eventually build is one you already know people want. If you’d rather run this loop with mentors and a room full of other founders, you can apply for free and only pay tuition if you get in.
But you don’t need any of that to start today. Pick your outcome. Find three strangers. Charge a little. Deliver it by hand this week, and let what happens tell you the truth before you write a single line of code.