Lean Startup: build, measure, learn.
Do I need to waste months of works to learn what my customers want?
This was the question that Eric Ries did after 6 months working on a project and see how it was useless. His problem was a classic one, develop a product just based on assumptions.
A startup is an idea, and it can be applied to create a new business or inside an enterprise could be a new section in the business. Unlike other business development techniques, it doesn’t have a huge study of the market, basically because what are we going to offer to that market, does not exist.
And the second thing to know, after understand what is a startup is how commerce works. For example, you suppose that a new product can solve a specific problem, make a market research, develop the product and put in on the market, expecting to see how customers react to the new product. That is the classic entrepreneur way, but in the real life for small business, for ideas, for startups, it is really expensive, and let’s be honest doesn't matter how much you study your market, the things can go wrong.
The reason that many startups fail is that they act convinced in how the customer will react to their product, expend a lot of time and money in their first idea. But at the moment to go outside, there is no market, there is not a user who will use their product.
Now what Eric Ries propose us in his book is to develop an MVP, a Minimum Viable Product, to test our hypotheses, learn about it and then apply it to a new MVP, after a while, we will get the real product.
So the Lean Startup is the methodology that helps entrepreneurs to avoid failure, or at least to take the failure in the cheapest way.
The essence of the lean startup is:
- A minimum viable product.
- Customers reactions
- Is your idea validated? or you need to adapt(pivot)?
And repeat that until your customer sends you a signal that your product is really covering their needs.
Now, the key part of the Lean Startup is how it validates the hypothesis, which is for the most part design I test. And the best way to do that is having a clear idea about what are your main assumptions that will determine your success. Once you got that, the next step is to find the cheapest and faster way to test them.
For example, I have an idea about a startup where “busy people” with pets, find “free people” to take a walk their dogs. We are going to call it Pet Walk.
The main assumptions are:
busy people will leave their dogs into the hand of a complete stranger.
For what I am going to pay, for time walking or for a distance walking.
How do I will search for a “walker” (the guy who will take my dog for a walk)?
the closer one to my location
the guy with most stars.
And yes, like I said before, the cheapest and faster way to test those hypotheses is developing an MVP. And have it clear that your developers have to be a focus in answer your hypothesis. For example:
Let’s make a simple MVP for testing our first assumption; There are many ways, everybody knows how dropbox tested their product, with a video explaining all the idea, was a complete success, we can do that, but in this case, I would prefer a closer relationship with the “busy people”. Unlike with Dropbox, I will find a different kind of possible clients, if I made a generic video it will be the same for everybody. And well, maybe people under 20 years old with dogs will not be interested, but professionals over 30 years old with no time will love my app. So in this particular case, I would not make a video, but neither will I start to develop the app. A social media campaign, with a very good selection of the market, a landing page explaining the idea for the different kind of clients, and with a form at the end of getting some extra feedback could be enough. And that is my MVP.
Now, the other two assumptions, are more complicated, maybe a form will not be enough. So, we need to develop something basic for the two options, I mean something really basic, something that just gives us the option to choose between a distance or a time walking, it doesn't mean give a detailed GPS system that will locate the walker, no!. We need something more basic, something before that, something before goes and expend resources and time of development. And is the same for the third case, develop something simple for rating users or for locale someone, maybe just with the address will enough forget about the GPS, this is a simple MVP.
And measure reactions, that is why is so important to know which is our hypothesis and the relation with your MVP.
Now the good part, what about if my hypothesis was wrong, and the user simply is afraid to leave his dogs in the hands of a stranger. Well is time to pivot… or die.
Pivot in the Lean Startup is a change in your idea, look for another way to do the same, based on another hypothesis, validated in other MVP. Sometimes there is more than one way to do the same, so it is easy to find a solution, also sometimes there are no more ways, or there is no more resource to keep developing MVP, so there is where the Startup ends.
But don’t feel bad about that, if you have enough resources for test the other hypothesis and go ahead. That is a startup, sometimes things do not turn out as we planned, and when that happens it is smart to have invested not too much.
So, when people talk about a startup, the first thing that I think is about how much risk it could have. It is not too much compatible with the predictive methodologies like PMI, but with a very well plan you can go ahead modifying the main idea until you can satisfy your customer's needs.