The Short Answer
What are some examples of a minimum viable product?
Dropbox tested demand with a demo video; Zappos sold shoes online before holding stock; Airbnb's founders hosted guests in their own home; Buffer used a landing page with a pricing step; Product Hunt started as an email list; Mondo, now Monzo, launched a prepaid card before getting a banking licence; and Superhuman measured fit on a live product. Each tested one question as cheaply as possible.
Eric Ries defined the MVP as the version of a product that gets the most validated learning with the least effort. The examples below show how differently that can look. Only some of them involved building software at all.
Seven MVP Examples
Filter by type to see which approaches have been used for the kind of question you need to answer.
Interactive gallery
Filter the examples by type of MVP
Each example links to the founder’s or company’s own account, or an interview with them.
Dropbox2008Demo video
Test: A short video showing how file syncing would work, aimed at early adopters, before the product was widely available.
Result: Founder Drew Houston told TechCrunch that 75,000 people joined the waiting list within 24 hours, against an expected 15,000 at most. TechCrunch interview
Zappos1999Manual behind the scenes
Test: Photograph shoes in local stores, list them online, and buy each pair from the store only after a customer ordered it.
Result: It showed people would buy shoes online before the company held any stock. Founder Nick Swinmurn describes the model in his own words in Fortune. Fortune, 2012
Airbnb2007Concierge
Test: The founders rented space in their own San Francisco home to visitors and hosted them personally.
Result: Airbnb dates its start to October 2007, when two hosts welcomed three guests, under the name Airbed & Breakfast. Airbnb newsroom
Buffer2010Landing page
Test: A two-page site explaining the product, then a pricing page added in between to see whether visitors would click on paid plans.
Result: Enough visitors clicked paid plans to justify building. The first version took seven weeks of evenings and weekends, and the first paying customer arrived within four days of launch. Buffer founder’s account
Product Hunt2013No-code
Test: A daily email of new products, set up with an existing link-sharing tool instead of a website.
Result: Founder Ryan Hoover told TechCrunch he created it in 20 minutes. It reached around 200 subscribers, and the positive response led to building the website. TechCrunch
Mondo (now Monzo)2015Partner-built
Test: Before it had a banking licence, the company offered a prepaid debit card and an iPhone app, labelled “Alpha”, to a first wave of 500 testers.
Result: It planned 3,000 limited-edition Alpha cards over about three months while applying for its licence, testing the app with real spending. Monzo blog, Oct 2015
Superhuman2017Live product, measured
Test: A live email product measured with one survey question: how would you feel if you could no longer use it?
Result: Its score rose from 22% to 58% “very disappointed” after it rebuilt its roadmap around the users who valued it most. First Round Review
What the Examples Have in Common
- One question each. Dropbox asked whether people wanted file syncing. Zappos asked whether people would buy shoes without trying them on. None tried to test everything at once.
- Real behaviour, not opinions. Sign-ups, clicks on a paid plan, a booked stay, money spent on a card. Buffer went further than most: its pricing page tested willingness to pay before a line of the product existed.
- The cheapest thing that answered the question. Product Hunt's email list took minutes. Zappos had no warehouse. Airbnb used the founders' own flat.
- Built properly once the answer was yes. Buffer spent seven weeks building a working product after its landing page test. Monzo built a real app on a partner's card. The shortcut was for the question, not for the product.
Which Type Fits Your Idea
| Type | Best for testing | Weakness | Example |
|---|---|---|---|
| Demo video | Whether people want an idea that is hard to explain | Interest is not usage | Dropbox |
| Landing page | Demand and willingness to pay | Clicks are not purchases | Buffer |
| Manual behind the scenes | Whether customers will buy, before automating | Does not scale; hides real costs | Zappos |
| Concierge | What customers value, learned by serving them yourself | Very small numbers | Airbnb |
| No-code | Whether people come back, using existing tools | Limited by the tools | Product Hunt |
| Partner-built | Regulated or infrastructure-heavy products | Partner fees and rules | Mondo |
| Working software MVP | Whether people keep using and paying | Costs more to build | Superhuman (measured) |
If the question is about demand, start at the top of the table. If it is about whether people keep using the product, you will need working software. Our guide to MVP vs prototype vs proof of concept includes a tool for deciding which you need next.
Where Shortcut MVPs Stop Working
Videos, landing pages and manual services test interest. They cannot tell you whether people will use a product every week, whether it holds up under real data, or whether your costs work at scale. That is when a working software MVP becomes the cheapest way to learn, and when it needs to be built well enough to keep.
At that point, the questions become scope, time and cost: how to scope an MVP, how long it takes, and what a $15k MVP includes vs a $50k quote. For a fixed-price build, see Loomstrat's MVP development.
How These Stories Get Misread
- “Dropbox launched with just a video.” The video tested demand. The product still had to be built, and syncing files reliably is hard engineering.
- “An MVP should be embarrassing.” It should be small, not broken. Buffer's founder called his first version rough around the edges, but people paid for it within four days.
- “Manual MVPs are free.” Zappos bought shoes at retail to fulfil orders. Manual work hides costs that appear when you automate.
- “Regulated products cannot have an MVP.” Mondo shows the usual route: launch on a partner's regulated infrastructure while your own authorisation is in progress. See fintech MVP development in the UK.
Frequently Asked Questions
What is a good example of an MVP?
Buffer is one of the clearest. Its founder built a two-page site, then added a pricing page to see whether visitors would click paid plans. Enough did, so he built the product in seven weeks, and the first paying customer arrived within four days of launch.
Was the Dropbox MVP just a video?
The demand test was a video. Founder Drew Houston told TechCrunch that 75,000 people joined the waiting list within 24 hours. The product itself still had to be built.
What was Airbnb’s MVP?
The founders rented out space in their own San Francisco home. Airbnb dates its start to October 2007, when two hosts welcomed three guests under the name Airbed & Breakfast.
What is a Wizard of Oz MVP?
One where the customer sees a working product, but the work behind it is done manually. Zappos is a well-known example: shoes were photographed in local stores and bought from them only after a customer ordered.
Can a fintech start with an MVP?
Yes, usually on a partner’s regulated infrastructure. Mondo, now Monzo, offered a prepaid debit card and app to 500 alpha testers in 2015 while it applied for its banking licence.
Do MVPs have to be software?
No. Videos, landing pages, email lists and manual services have all been used to test demand. Testing whether people keep using or paying for a product usually needs working software.
The lesson from all seven is the same: decide what you need to learn, then build the smallest thing that teaches it to you with real customers.