How to Design Experiments to Validate Your Business Model

As an entrepreneur, if you have a business idea, sooner or later you will need to design experiments to validate your business model . It is a fact that when we start thinking about a project, all we have in our minds are hypotheses or ideas of how we believe the market works or how it will react to our product or service.

Therefore, if we start from the point where we already have our business model, either in a Business model canvas or in a Lean Canvas , we still have a long way to go to see if our business looks profitable or not.

How to propose experiments to validate our business model

Imagine that you have your complete canvas, with the hypotheses of your idea. For example, in this case I suppose that I want to develop a website to sell Stevia or any other product.

One of the most critical hypotheses that we must validate in our business model will be, for example, knowing if the chosen customer segment is really real or not . Namely:

  • If it is a segment that spends money on this type of product.
  • If the money they invest makes the business profitable
  • If it is a wide enough market to scale my company.

To achieve this I lay out the canvas above in 3 different layers.

The layers of the business model that help me to validate the hypotheses of my business

First layer. The business model canvas

It’s the one we have upstairs. That is, the lean canvas or the complete business model canvas.

Second layer. The most critical hypotheses to Validate your Business Model.

The second layer would be the same canvas but leaving only the hypotheses that I want to test. Because of everything we have written in the post its, there will be things that we do not need to check or that fall by their own weight. However, others will be hypotheses so important that they can ruin our business.

In the example above, for example, if I want to launch a platform on the Internet, one of the critical hypotheses will be that this market segment uses the Internet to buy these types of products.

If they don’t use the Internet, my business as I have it planned right now falls to me.

Therefore, in this second layer, the objective is to capture only the most critical hypotheses of that business that I have in mind.

In the previous case, as an example, we could limit ourselves to leaving the canvas with these hypotheses:

  • The target customer is a young 35-year-old woman who takes care of her physique and appearance and wants to be healthy.
  • Customers value the delivery of recipes, courses or exclusive shipments.
  • The online sale of this product leaves enough margin or not.
  • The target customer sees the value proposition as real or not.

Third layer. Design experiments for each hypothesis

Yes, what you hear. It is very likely that each hypothesis requires an experiment. It is a job that requires a lot of effort and this is where most entrepreneurs fail.

The Lean Startup methodology is clear on this point. If we do not get validated knowledge , the effort is useless. It is necessary to propose experiments that allow us to learn.

That is, from the previous hypotheses that we have selected, it would now be a matter of proposing an experiment for each one of them.

The first, without a doubt, is always to get to know our target client. In this case, the tool we use is the problem interview, because it allows us to learn a lot with relatively little effort.

Another example of an experiment that occurs to me is, for example, to validate the hypothesis of whether customers value the sending of recipes, content or ebooks, do the following:

  • Create a landing page with a registration form. In that landing, the value proposition is precisely to register to receive information, content and recipes about stevia.
  • Create an ad campaign on Facebook, to attract traffic to the landing page. The great advantage of Facebook is that you can very well segment the target audience to direct your campaign to. For example, 35-year-old women who like sports, healthy living and who are in a certain country.
  • Set a goal. For example, registering 3% of the traffic derived from the landing page.
  • Delivery of content and products. With the emails of the people who have registered, send offers of products and recipes and check how many people open the emails or buy the offer you send them. Here too we should set ourselves a goal. For example, they buy 1% of the people to whom you have sent the offer. If you send content and recipes, for example, 50% open the emails and 25% click on the content.

If the experiments are positive, the hypothesis is validated and therefore, you already know that these types of people value what you are going to offer on your website.

But is designing and executing the experiments everything to validate your business model?


It is necessary to do it quickly. The more laps in the shortest time we give to this circle, the better. More knowledge of the market we will be acquiring.

When possible, it is always better to put a metric that allows us to say if the experiment has been ok or not ok. As far as possible, the limit should be as clear as possible, so as not to give rise to confusion that makes us go down the wrong path or make us reconsider a business that could have worked.

How long does it take to run an experiment to draw conclusions?

Another fundamental aspect when developing your experiments is the time needed to draw conclusions. For example:

How many problem interviews should you do until you know if the client you have raised is correct or not?

Without going into statistical issues, you should continue interviewing people who meet the profile you have defined until you see that you do not obtain new relevant information. There is no exact number. In the Scrollthenews projects, sometimes 10 has been enough for us and other times we have carried out more than 50.

What do I want to learn from my experiment?

The million dollar question. It is always necessary to learn something from each experiment, if not, we have planned it wrong. In the process I have described, we always start with a hypothesis that we want to test. Therefore, learning will be related to that hypothesis and the experiment should give us a clear result.

Both offline and online tests and experiments must always be associated with a metric that allows us to know and evaluate the result to end up validating your business model or not.

Sometimes we entrepreneurs believe that just by laying out our business idea on a canvas or in a business plan, everything is done. But reality, the market and customers are responsible for putting everyone in their place. Only those who know the market very well, because they have thoroughly researched, experienced and worked on it, are the ones who have the chance of developing a profitable business.

Most entrepreneurs who fail do so in this part, in testing their business hypotheses. Therefore, it is key to have a method that helps us and guides us.

Leave a Comment