How and why to work hypothesis-driven when building (micro-mobility pony sharing) ventures
When was the last time you made a hypothesis? I would guess It’s been shorter than you might think. We constantly make basic hypotheses or assumptions in our heads. “The layer of ice is definitely thick enough to skate on” or “this dog looks cute, it won’t bite” are just some examples. We need these assumptions to handle the complexity of the world we live in — in our private everyday life but also in almost every work we do, no matter what profession.
Hypotheses are basically suppositions based on limited evidence as a starting point for further investigation. That doesn’t only sound very scientific, hypothesis-driven thinking is the standard approach in science to create knowledge and the starting point for research. And as this has been proved to be a good way to enter unknown areas, why not apply it in other fields — for example… building ventures?
It’s all about handling uncertainty
Following a hypothesis-driven approach isn’t by far anything new to the entrepreneurial world. At least since Eric Ries published his book “Lean Startup”, hypotheses made their way into startups all over the world. And their main function is as easy as that: reduce risks. In our context risk mainly means: wasting time and resources building products that no one wants. In order to avoid that, we need to reduce uncertainties — simply because there are far more unknowns and uncertainties at the beginning of every entrepreneurial journey than known facts.
Using hypotheses in venture building
At Sparrow Ventures our goal is to create value for the Migros group as a whole. We do this by investing in existing startups but also through building our own ventures that are created and run by a multidisciplinary team of experts. The latter typically involves building several startups from scratch in parallel with a systematic approach. For this, we rely on proven methodologies to validate innovative business models with speed, and following a hypothesis-driven approach is one of the most crucial.
Let’s jump into it and get more specific, on why this is the case. Whatever topic we are looking at and no matter in which phase of our venture building process we are in — from topic validation over prototyping to MVP — the most important question we constantly ask ourselves is: What are the most critical hypotheses at this point in time?
While there are many examples from our daily work that we could use, let’s take something fun with a fictitious example and assume that we want to enter the field of building a micro-mobility pony sharing service (credits for the idea, recommended watching for entertainment). So before we even start doing anything, we would formulate basic hypotheses:
These hypotheses can touch various aspects of a potential business, such as
- Customer problems
- Market size and trends
- Financials / Unit economics
- Viability of a solution
- Potential channels to reach customers
In general, our hypotheses are high-level at the beginning of the process and usually become more granular as we gain more knowledge and insights in the process of developing a venture further. Something important to keep in mind here is that these hypotheses are only valuable if they are falsifiable, meaning they can be rejected through tests, research and experiments. In this regard, including quantitative metrics where possible makes life easier.
Structure and prioritize with a Hypotheses Log
To allocate time and resources effectively, we always try to focus on the most critical hypothesis at any point in time. Therefore, it is a crucial next step to prioritize these formulated hypotheses in a meaningful way. We do so by basically evaluating all hypotheses along two dimensions: uncertainty and the potential impact.
Uncertainty describes our current estimation about how certain or confident we are that a hypothesis is true or can be validated. We use a basic scale from 1–5 with 5 having a very high uncertainty and (almost) no insights on the topic so far.
Potential Impact is an estimation of how the (potential) venture will be impacted by a specific hypothesis not being true or falsified. Also here we use a scale from 1–5 with 5 being a very high impact.
Having these dimensions is making it now easy to prioritize hypotheses based on what we call a risk score, which is simply calculated by multiplying uncertainty and potential impact. The higher the risk score, the higher the priority on validating a hypothesis should be.
In our case of a potential pony ride micro-mobility service, a section of the hypolog as we call it could look like that:
By that simple but effective approach, we make sure to always focus on the most critical hypotheses. An advantage of this dynamic “mechanism” is that as soon as you gain more insights on a topic and the uncertainty decreases (and therefore also the risk score), other hypotheses automatically go up the hypolog and move into the focus.
In order to not get lost and just jumping on the first best thing that comes to mind, formulating and structuring hypotheses helps a lot to stay focused. Therefore, the hypolog is and should be a truly living document that is constantly being updated and worked on by the venture team, enriched with responsibilities, details on the current status, and how we plan to test specific hypotheses.
Formulating the hypotheses is only the beginning
Talking about testing or validating hypotheses, this is where the actual work starts. By formulating hypotheses we just laid the groundwork. In this case, testing would involve in a next step qualitative and quantitative work such as
- talking with potential customers about their problems with current mobility solutions
- calculating unit economics by talking to market experts
- estimating potential market size through research
In this way, step by step hypotheses will be validated or falsified and more insights are gained to reduce uncertainty. For us at Sparrow Ventures, this also means to make sure to only develop products and services that really solve a customer problem and contribute to a sustainable improvement of people’s quality of life in Switzerland. However, diving deeper into the mechanics and different methods of testing hypotheses is up for another blog post. For the time being, whenever you hear yourself saying that the layer of ice is definitely thick enough to skate on, remember the approach described and quickly estimate the risk score of your hypothesis — you may want to reassess your decision.
This blog post was written by Leon, who is a Venture Architect in our team.