In his seminal book, The Lean Startup, Eric Ries defines the MVP as a version of the entrepreneur’s product which enables a full turn of the Build-Measure-Learn cycle with minimum effort and the least amount of development time. The purpose of the MVP is to test one hypothesis.
Every startup business is predicated on a number of assumptions. For example, typical assumptions might be:
- Customers need our product.
- Customers will value this particular feature.
- Customers will find our product more convenient than X.
- Customers will pay Y for the value we provide.
- Customers will be able to install and use the product without help.
Each of these is a separate and testable hypothesis. Ideally, we would produce a separate MVP for each hypothesis and test it with a different small group of customers. You would start with the biggest and most fundamental assumption and ship a product which tests that assumption. The product will lack many fine features that you intend to include in the final product but that does not matter. What matters is that it enables you to get an answer to the key question – is this assumption valid? If the answer is negative then you have gained an invaluable piece of learning. There is no point in carrying on to build the full function product. You must rethink your whole plan.
The point of the MVP is to learn something important quickly. Sometimes the MVP is so minimal that it is hardly a product at all. When Drew Houston founded Dropbox he wanted to test the hypotheses that people had a real problem with file synchronization and would use a cloud based solution. Building a working prototype was very difficult so he did something much simpler. He made a video. The video showed how the product would work; Drew narrated the story of how a user could store and retrieve his files. The video worked and drove a tremendous number of enquiries. It was the MVP which validated the assumptions.
Another minimal approach is what Ries calls a smoke test. Customers are given the opportunity to pre-order a product which does not yet exist. The smoke test measures just one thing – whether customers are interested in trying the product. Nonetheless this is a highly useful piece of input.
A more ingenious approach is the so-called Wizard of Oz test. Here customers think they are interacting with the full automated product but in fact a small team of people behind the scenes carry out all the functions of the product or service. This was the method used by Max Ventilla and Damon Horowitz when they were developing Aardvark, a highly sophisticated web service which could produce answers to complex subjective questions. Using humans to replicate the service was costly but allowed the key assumptions to be tested with a small number of customers. Eventually Aardvark was sold to Google for a reported $50m.
In 2004 Mark Zuckerberg, Dustin Moskowitz and Chris Hughes raised $500,000 in venture capital for their fledgling company Facebook. At the time they had 150,000 users but virtually no revenue. A year later they raised a further $12.7m. How could they raise so much money so early? Crucially they could demonstrate to investors that they had validated two key assumptions about their business model. More than half the users came back to the site every day. This proved that customers found the product valuable (what Ries calls the value hypothesis). Secondly, their rate of growth was staggering – within a month of launch at Harvard almost three-quarters of undergraduates were using it. This validated the growth hypothesis.
The MVP is often a low-cost, low quality product but the accelerated learning it delivers can mean the difference between life and death for the startup. Craigslist and Groupon both started with small low-quality product offerings. Stop believing your own publicity. Instead, test your biggest and riskiest assumptions quickly with a Minimum Viable Product.