Building software first is the most expensive way to discover that nobody wants it. Before you spend months and lakhs on an app, you can test whether real people will actually pay by delivering the promise entirely by hand. Two lean techniques, the concierge MVP and the Wizard of Oz MVP, let you sell and serve customers with almost no code, so you learn from paying users in days instead of quarters.
Two ways to fake a product before you build it
The concierge MVP: do it manually, in the open
In a concierge MVP you deliver the service by hand, and the customer knows a human is doing the work. You are the "software." Eric Ries popularised this idea in The Lean Startup using Food on the Table. Founder Manuel Rosso wanted to build an app that generated personalised weekly meal plans and shopping lists based on local grocery deals. Instead of coding, he visited a single customer's home in Austin, asked what her family liked to eat, checked what was on sale, and hand-built her meal plan and shopping list every week for about ten dollars. Serving one person by hand taught the team what customers truly valued, which turned out to be diet and allergy preferences more than coupon savings, before a single line of product code was written.
The Wizard of Oz MVP: automated on the outside, human on the inside
A Wizard of Oz MVP looks like a finished, automated product to the customer, but humans do the work behind the curtain. The classic example is Zappos. In 1999 Nick Swinmurn wanted to test whether people would buy shoes online. Rather than buy inventory or build warehouse logistics, he photographed shoes at local stores, posted them on a simple website, and when an order came in he drove back to the store, bought the pair at retail, and shipped it himself. He often lost money on each sale, but he proved real demand cheaply. Zappos grew into a major retailer and was later acquired by Amazon.
Concierge versus Wizard of Oz
The difference is transparency. Concierge is openly manual and high-touch, best for learning deeply from a handful of users. Wizard of Oz hides the manual effort behind a real-looking front end, best for testing whether people will click, order, and pay as if the product already existed.
Why this matters for a first-time founder
Manual MVPs flip the risky order of things: you get real behaviour and real money before you commit engineers, capital, or your savings. You also learn the messy operational details that no survey or pitch deck ever reveals.
How to run one in India this week
- Pick one sharp promise and 5 to 10 real target customers, not friends and family.
- Deliver it by hand. Use WhatsApp, a Google Form, and a simple landing page or Instagram page as your "front end."
- Charge real money. A paid yes is the only trustworthy signal. Collect through UPI so payment friction is almost zero.
- Stay light on formalities. You do not need to incorporate a company on the MCA portal via SPICe+, or obtain DPIIT Startup India recognition, just to run a paid pilot. Those steps come once demand is proven. A small, local pilot will usually stay well below the GST registration turnover thresholds, which are Rs 20 lakh a year for services and Rs 40 lakh for goods in most states, and Rs 10 lakh and Rs 20 lakh respectively in special category states. Register for GST once you cross the threshold or once your model otherwise requires it.
What to measure, and what to avoid
Track conversion, meaning how many prospects actually paid, plus repeat usage and the requests people make for things you did not offer. Avoid two traps: running the test forever instead of deciding, and hiding that the service is manual when a customer's money, safety, or legal rights depend on knowing. Once the manual version shows clear paying demand, automate only the parts customers proved they value.