What are startup ecosystems?
What’s a (startup) ecosystem?
I ended my third article on the ideal startup support system with the answer to the question of the ideal startup system: it should be an ecosystem. I’ll use this article to explain what an ecosystem actually is and how it works, but let me start with recapping why an ecosystem is the solution: Because a startup is an ever changing entity it is very hard to support the startup with a fixed structure as a program for longer periods of time. This also counts for mentors and even for VCs. They all have their expiration date. The moment when a mentor is out of experiences to share which are applicable to the startups challenges or when a VC can’t afford the next round of financing the support is done for. You could solve this by creating a flexible program, or a flexible fund, or, as a mentor, to create new experiences just as fast as the startup needs them. It is, however, much easier to accept that the support will end at a certain point, and someone else will take over.
Another mentor, more experienced with the next set of challenges, a new VC fund with bigger pockets and a new program. When you accept the fact that a fixed program, or a fixed coach only has a temporary role your role as a support organization changes. It becomes two fold: making sure the people, organisations and programs are there for every phase, and to make sure the handover between the different actors goes smoothly. Orchestrating this web of actors (founders, entrepreneurs, mentors, universities, incubators, VCs, etc) and factors (financial capital, human capital, intellectual capital, etc) is the key to creating and building the ecosystem which is needed for proper, long-time, value-creating support.
Why still so many programs?
But why then do we still see so many organisations solely focused on programs and coaching? Which is, by the way, not necessarily wrong to do, you just have a limited value in the whole journey. I’m not sure, but my guess is that If you accept the temporary value of a fixed structure, you need more coaches, trainers, mentors, programs, VCs etc than ever before. Perhaps organisations don’t have the budget, the time, or the energy to create such a system. Besides that he handover is very hard to orchestrate. You need an information sharing system and you need more openness and trust within the entire ecosystem.
This brings us to the problem with ecosystems: they are hard to create, build and manage. They take more time than we’re used to taking (and receiving). They need more budget for a longer period of time. Results will not be immediately there. Brad Feld talked about 20 years timelines, and it is one of the reasons he thinks that ecosystems can not be created and managed by support organisations, but you need the (local) entrepreneurs to lead it. I disagree with him on this, but I still see the difficulty of these timelines when we live in an era where local and national government tends to think in 4 year timeframes, and even VC funds horizons stop at 10 years.
Anyhow, a lot of support organisations stick to the program(s) they have and know how to run, without claiming their role in the ecosystem and thereby limiting their value and influence on the entire system. I hope this chapter will change this (a little).
Ecosystems as complex systems
Imagine a shared event space, or some shared office space. A spot where a few startups come together. YES!Delft started in an old campus building for instance. Someone creates a little budget, you get a shared kitchen, a shared reception and a shared loading dock. Another one manages to get his or her hands on some tools for a shared workshop. Someone’s neighbour walks in, he made a successful exit and shares his experiences during a coffee talk. And hey, before you know it, it starts to look like an ecosystem.
There would be multiple VCs active who could (jointly) organize programs focused on investor readiness. This saves them time as they will meet the startups very early-on and they can build relationships with the startups before it is really necessary. Pretty soon a bank and a law office would join as they want to meet their future customers as soon as possible as well. The startups will attract interns from the university, some will join the companies as employees, others will become ambassadors for the ecosystem. And slowly the ecosystem evolves together with the startups.
And this is also where the trouble starts: as more parts are added to the shared office space. As more startups join, as more people walk-in and start to share experiences the simple system of a shared kitchen and loading dock starts to change into something more complex. And although complex systems are everywhere around us (forests, coral reefs, traffic, a human cell, your family) we still have a hard time grasping complexity. Complexity is not just more of the same complicated stuff, it’s a whole different ball game and our education did not prepare us to deal with it. I’ll explain with the help of a Cynefin framework:
On the axis are the “how” and the “what”, all the way on your left you can find the clear area. You know what to do, and how to do it. Tying your shoe laces is on the clear tasks in your life. If someone taught you how to do it drilling a hole or hammering in a nail are also clear tasks. Up next is the complicated stuff. Here comes education into play. These are excel sheets, financial modelling, mathematics, physics, language, etc. If you put something in the model, something comes out. There are clear relationships between what you do and what the results are.
The issues can be hard, but you mostly know what to do, and how to do it. That’s why you can teach them to others, and learn about them in a book. In that sense complicated stuff is just a big bunch of clear tasks put together, and you have to perform them in the right order. But what happens when you’re not sure anymore what to do? And how to do it? That’s where complexity starts. It will not work to gather a bunch of clear tasks and perform them in the right order, because you’re not even sure what the tasks should be! You’re in a whole new area, which is closer to “chaos” than to “clear”. An area where your high school education did not prepare you for, yet your instinct still tells you it is more of the same you know so well and it tries to handle complexity with more of the complicated stuff. What happens when you use your complicated mindset in your complex world?
You try to predict what happens, but you’re wrong most of the time. You have trouble making estimations, especially on how long things will take. You’re only seeing one side of the puzzle, so you’re surprised by the outcome of meetings, decisions were made that you didn’t see coming, etc etc. Startups, especially the ones in new industries, or who use new technologies, are in the complex zone. One of the ways to deal with the uncertainty in a startup is the lean startup method of Eris Ries, which are also applicable to ecosystems!
Because ecosystems are so different from what we are used to, it is useful to use models to talk about ecosystems and to create a common language and understanding of the ecosystem.
Ecosystem Models
Actors & Factors
The easiest model to describe an ecosystem is the actor/factor model. It is basically a list of all the relevant people, organizations and resources involved, divided into actors (people and organizations) and factors (resources and conditions) in a visual representation. I used the image from Brad Felds The Startup Community Way, and altered it slightly to make my points come across.
Three things pop out in my opinion: the ecosystem is centered around entrepreneurs, not around any of the other actors. (Support) organizations tend to put themselves in the middle when describing “their” ecosystem. This is almost never the case. A forest does not revolve around the forestry organization, just as the coral reef does not revolve around the organisation trying to protect it. The essence of a support organization is that whatever they support is in the center of their purpose, not themselves. As mentioned before: the startup sometimes wanders out of view as the customer. These models help to put them back where they belong.
The second thing which stands out is that everyone is connected through the cultural capital, and the ecosystem is open (notice how there is no boundary around the cultural capital?). It is not locked by some artificial boundary like a municipality or an industry. This means that if you share the same culture you can be in, and making connections to other actors should be fairly easy
The third thing is the mixture between individuals (employees, serial entrepreneurs, founders, angels, mentors, trainers) and organisations (universities, large corporates, service providers). This makes for an interesting mix as individuals are much more fluid than organizations. They are faster to join ecosystems, but are also easier out when they do not find what they are looking for.
Network models
What the actor/factor model lacks is a focus on the interconnections between the different actors. Are some actors more important than others? Are there so-called supernodes that create more connections than others? How often do actors actually contact each other? You can solve this by a network model, where you get visualisations similar as the one below. If you then label the nodes as actors and label the connections with factors you get a quantified actor/factor model. The problem with these kinds of models is that they take an enormous amount of time to generate and are still lacking information on the interaction. It is of course not only about the frequency of the interactions, but mostly about the quality of interactions. What you want therefore is not so much counting the interactions, as well as an indicator of the quality of the interaction.
An example from YES!Delft
At YES!Delft we generated our own network model by introducing a (communication) platform which could be used by our mentors, trainers, startups, and service providers. We could count the amount of interactions, and both startups and the other actors could rate every interaction they had with each other (both the startup and the coach received a rating). As an administrator you could check all the ratings and quickly “fish” out the ones which had low(er) ratings for a certain period of time.
Logical Models
Creating a logical model of the ecosystem is a very good exercise for every stakeholder in the ecosystem. What are your expectations and goals of the ecosystem? What are your goals? What would you like to achieve within the ecosystem? You start the model either with your resources, or the activities you already have, or with the problem or the goal you have in mind. When filling it, it works the same as the business model canvas, in that you keep iterating the different blocks until you have a model which you are happy with. It will show you the pain points of your ecosystem, or where a number of stakeholders are not aligned (they could see different problems, or have different goals, or do not see the relationship between the outcome and the activities. In that sense a logical model is much more a discussion tool than a model in itself. This also means that you shouldn’t worry too much about the details in the model, it is much more about the perceived relationships between the different parts.
Challenges with ecosystems
By now, you’re probably already thinking about some ecosystems you know: Silicon Valley of course, perhaps Bangalore, or the area around Shanghai and Shenzhen? Ecosystems already exist, they seem to come out of nowhere and without guiding and managing them, so what’s the issue here? Why these articles in the first place? Or stated otherwise: Instead of competing programs, competing VCs, competing incubators and competing individuals why are there only so few successful ecosystems?
The problem is: The successful ones have evolved spontaneously instead of by design, and for every successful one there are many failed ones, or ones that only lasted for a few decades (one of the more recent ones is the total crash of the ecosystem around Detroit). Our success rate is just not that high when it comes to ecosystems. The most important reason for this is that ecosystems are complex systems, and we’re just not good at understanding complex systems (similar as to why we still can’t seem to understand exponential growth).
As complex systems, the value of ecosystems is not in the parts (incubator, programs, startups, service providers, large corporates, etc) but in the interaction of the parts. We clearly know how to optimise the parts (create the best program, create the best legal firm, create the best university, create the largest corporate), but it is yet unclear how to design and manage for optimal interaction between the parts.
An example from YES!Delft
I have seen this issue myself in my years at YES!Delft as well. As a support organization you might be very good at optimising the part (the startup), but optimising the ecosystem is a whole different issue. For example: we tell founders to focus and spend time on their startup (optimise for the part) and that this focus is one of the key elements of success (for their part). On the other hand the ecosystem only works if founders also focus on the interaction between the parts (i.e. to take the time to answer a few hard questions from your neighbouring startup).
In practice we saw this with hosting events like peer 2 peer sessions or unconferences. They were always very highly rated afterwards, but we had a hard time convincing people they should not spend time on their business for an afternoon and instead come and listen to their fellow entrepreneurs.
One of the reasons is probably the (very) slow positive feedback loop of spending time on the interactions. Spending an afternoon on your company gives immediate results, spending time talking to others has an output which is much less clear, in other words: you are inclined to spend time on the part, instead of on the connections.
Who owns the ecosystem?
Another issue is the vagueness of the ownership of the ecosystem. The ownership of the parts is clear: the founders own the startup (at least a part of it), the VC also owns a part of the startup and their own fund (well, also a part of it..), and the incubator is managed by a board. The problem is not in the parts of the ecosystem, the problem is: who owns the interaction? Who is (and feels!) responsible if VCs start competing instead of cooperating? Who will take action if neighbouring startups stop talking to each other? Who creates some insight in the ecosystem, and tells you who is who? This list of activities, as you probably guessed by now, is endless. But perhaps this is the wrong question to ask, and ownership is the wrong issue. Maybe ownership doesn’t really matter, but it is about who possesses it (remember Peter Thiels explanation of ownership, possession and control?).
Who has possession over the ecosystem
Surprise, surprise. No one possesses the system. If change is the essence of a startup, then the lack of possession is the essence of an ecosystem. You might possess parts of the system (the incubator, or the service providers, or the large corporate), but you will never possess the interactions (which is where most of the value is). You might be able to steer the interactions in a certain way, you could monitor them, but no one possesses the right to communicate (or not communicate) with each other.
Think of it as a forest. If your goal is to get as much wood out of the forest as possible, you hire someone who cuts down trees regularly, who opens up areas so other trees can grow, etc. You might walk through the forest several times a year to check how things are going. Do you possess the forest? By no means. Do you have control over what happens? Can you push or pull it in a certain direction? Yes, but always to a certain extent, and you have to get used to the fact that every tree you cut will have unexpected results as well.
A startup ecosystem is no different from a dense forest or a coral reef. You are, with a lot of effort, able to push and pull it in a certain direction, but that’s about it.
So much to learn..
To summarize the above: ecosystems are the complex answer to the ever changing startups because the complex system will have all the different entities needed to solve every challenge and the system will transform together with the parts (the startups, the VCs, government bodies, etc). The problem is: how to design and maintain an ecosystem?
We still have a lot to learn on how to work with and in ecosystems, and experiences are not yet shared enough. And, perhaps we shouldn’t look to the industry for guidance, but look at how we manage other ecosystems in the world?
