How to test a business idea without wasting time and money?

Let’s think of a group of friends discussing a new business idea. An online platform that would allow individuals to rent their car for a few days or hours to other individuals visiting their area. The equivalent of what Airbnb offers, but for cars instead of homes.

Are there successful applications of this business model? A quick search for "peer-to-peer car sharing" will find examples such as Turo, which was launched in 2010 in the US and raised $500 million in investment, and Drivy, founded the same year in France and was acquired for $300 million, in April 2019, by Getaround. The model looks interesting. But is it appropriate to implement it in Greece for example - or any other European country- in 2023? How can we discover it? After all, how can you know if a model will work if you don't make the investment to try it, right?

The answer is no, many times it's not like that at all. For most business models like the example above, there is a lot we can do to measure the risk and avoid exposing ourselves, more than actually necessary. Over the last two decades, several methodological tools have been developed and disseminated to better manage risk in business and reduce wasted resources. One of these tools is the "lean startup".

Let's look for a moment at the example of Airbnb, which has been a pioneer of the "sharing economy". In March 2008, the founders of Airbnb created awebsite that allowed conference guests in various US cities, to find locals who would host them in their homes. The value proposition for travelers was the ability to stay at a low-cost lodge, but with a personalized hospitality experience that cheap hotels don't offer. For the hosts, the value proposition was an easy-to-get supplement to their income.

The model didn't work at first and within a few months it began to change, gradually approaching the current version that has allowed the company to have a $100 billion market capitalization. But the interesting thing is what preceded it in October 2007. Two of Airbnb's three founders had recently moved to San Francisco and were struggling to make ends meet. They didn't have enough money to pay the rent and were looking for ideas for quick income. They heard from friends that the hotels in the city were highly occupied and those who wanted to travel to the industrial design conference that was being held in the city, could not find cheap accommodation. So, they came up with the idea for a platform that would allow conference attendees to connect with local hosts. To test the idea, they thought of renting space in their own apartment.

They had no beds to spare, but they had an airbed and enough money to buy two more. They set up a website called "Airbed and breakfast" and rented the inflatable mattresses for $80 a night to three strangers.

With this experiment they reduced their uncertainty around the most critical assumption of the business model. The assumption that there are travelers who would sleep on an inflatable mattress, in the home of someone they met online, and would pay for it. The experiment cost them minimal time and money and gave them the confidence to go ahead.

Back to our example. The car-sharing business model has been tried before but not in the Greek market. There is a) a necessity/desire risk (is it attractive enough?), b) a viability risk (will it work with a sufficient profit margin), and c) a feasibility risk (e.g. are there any obstacles?).

How do we approach it? First, we start with the necessity/desire risk, for two reasons. Firstly, because it makes no sense to study whether something is practically feasible or theoretically viable, if it is not desired by thecustomer. Secondly, because if we focus on understanding the customer's needs, there is a good chance that we will discover other problems worth solving.

Action sequence:

Phase A- Data collection: talk to as many customers as possible, trying not to influence them with our idea but to find out their real priorities. We are mainly interested in understanding how they actually act. We want to not dwell on what they think they are doing or say they are doing. For example, answers to hypothetical questions such as "Would you rent your car to a stranger for 15 euros a day?" are not very useful. More valuable are questions such as: "When was the last time you borrowed your car?" "Have you ever refused to lend your car and for what reasons?"

Phase B- Identify and prioritise hypotheses: we list all the hypotheses that need to be valid for our model to work. E.g., the key assumption in the Airbnb model was that many travelers would not hesitate to pay to sleep in the home of an unknown individual. In the car-sharing model, it is perhaps a more important assumption that many would not hesitate to give their car to an individual they met over the internet. Each hypothesis should be scored on two dimensions, i) the degree of uncertainty and ii) the degree of impact. First, how much data do we have to support this hypothesis? Second, in the event that the hypothesis is shown not to be valid, what is the degree to which our business model is over turned? Do we start from scratch again, or do we make a small adjustment?

Phase C - Design and run experiments: we start from the assumptions that, if they were not valid, they would cause the biggest revision to the business model and are also the ones for which we have the least data. This is where the greatest risk lies. Depending on the hypothesis we design an experiment. For example, we contact 20 car owners who seem suitable, offering15 euros a day to take their vehicle and record how many will accept. We repeat until we find 10 owners and proceed by completing the rental and recording reactions. At a cost of 150 euros, we will have learned a lot about how our customers think and be ready to repeat the Build-Measure-Learn cycle that takes us from idea to test and from there to knowledge, gradually reducing our business risk.


This project has received funding from the European Union’s EIT HEI Initiative: Innovation Capacity Building for Higher Education under Grant Agreement No 10047.

By using this website, you automatically accept our use of cookies. Read more.