People join ventures because they like building new things, the journey into the unknown, the culture of "us against the rest of the world" – you name it. So why should startups plan their end?
Having an executable exit plan helps with prioritization and understanding which pathways have synergies / are mutually exclusive. Even if you are looking to build a self-sustaining business to keep full control of your company and stay independent, your investors are likely going to want at least some validation of the exit potential prior to investing. And for startups developing breakthrough technology in an area where marketing the product by themselves is not realistic (i. e. biotechnology, many medtech cases, an array of consumer products, etc.) - the exit strategy is even critical for survival altogether.
As I have recently been asked to weight in on a number of business plans, I realized that I wasn't aware of any useful framework to visualize exit strategies and make them communicable and understandable. Many startups I have worked with just put a number of exit targets in their business plans or reference some past exit success stories from their industry, but they hardly ever follow a well-defined process such as CANVAS for business model design when doing so.
Infographic: Mapping Potential Startup Exit Pathways
Game theory offers a useful background in how to optimize behavior in multi-stage games. An exit strategy is exactly that: A multi-stage game, where at different stages you will have different probabilities for getting to the next stage, resulting in an exit scenario. Once you define all potential pathways, you will more easily be able to pursue the one strategy that maximizes expected value.
This process in game theory is called backward induction. You start at the potential end-states, then define pathways, probabilities and intermediate stages that get you there. This allows you to prioritize the actions and deliverables today that are most likely going to lead you to your first-choice exit scenario 2, 3, or even 5 years down the road.
Sounds complicated? Then let's illustrate this with an example.
Example: Needle-free blood glucose measurement
Let's assume that you have developed a sensor to measure blood glucose levels without having to take a blood sample (btw: 7 promising technologies that can do just that are described here).
The following potential targets for an exit come to my mind:
Big Pharma / Insuline / Diabetes Care (i. e. Novo, Sanofi, ...)
Manufacturers of blood glucose monitoring devices (i. e, B. Braun, Beurer, ...)
Manufacturers of diabetes pumps (i. e. Medtronic, Ypsomed, ... )
Now here´s the catch-22: In order to be an attractive target for a pharma company, you may have to make some choices that make it unattractive for a medical device manufacturer to buy you, and vice versa (i. e. should you prioritize certain types of insulin in your clinical trial? How many resources should you invest into building your own QMS if an acquisition is likely to take place a few years before you have a commercial product? etc.). This is why laying out your pathways to an exit is so important: it allows you to map which choices have synergies, and which ones are mutually exclusive.
Start in the future, move towards today
Backward induction means you start at the end-states and then you move backwards towards the Now from there. For each potential end-state you should develop an idea on:
expected overall probability of success
expected lead time
required milestones passed
viable assets, viable liabilities, viable skillset
key activities in order to progress to next stage
(if you would like to keep things simple, just do items 1-4). Then you map each path in order to see where the nodes separate some paths from others, with eventually all paths leading to today. Overall, as with any business strategy framework, I would recommend to not make things complicated and be pragmatic. It´s unlikely you´ll need to plan for more than 5 scenarios. For an illustration of our blood-free glucose measurement example, please see below: