In looking to understand accelerators better, in my PhD I researched what a startup accelerator is by looking at the underlying social networks that form them. In the business I set up, EyeFocus Accelerator, I also experimented with how to build accelerators with an eye on the social network function driving them.

It may at first seem like an obvious statement, or not, that a startup accelerator is ultimately a social network, being a group of people linked together through social ties. However, accelerators are often seen primarily as business incubation programs, which happen also to be social networks. I would argue that they are primarily social networks, which when structured correctly achieve business incubation outcomes.

In this essay I will explain some of the social network dynamics in a good startup accelerator, and why these very specific dynamics support innovation and incubation. In doing so, I will also explain what accelerators are, along with the associated concepts like startups, mentors, and ecosystems.

What is an accelerator?

In 2005 Paul Graham, an American entrepreneur and investor, established the first accelerator program, Y Combinator. He decided to invest small sums of money in a cohort of tech entrepreneurs and support them as a group to develop their concepts. His intention was to create an efficiency by investing on the same terms into a cohort, and to be able to offer them the same support at one time. This was instead of acting like an angel investor and investing individually in each company, with separate terms and negotiations, and then supporting each company on an individual basis.

Since then, the concept of startup accelerators has spread from the US to Europe, and become a global phenomenon. Within academia, the study of accelerators is relatively new, and therefore lacks large data sets or longitudinal studies. This makes it difficult to evaluate definitively whether they work or not, and to reach a single definition of an accelerator. In studying accelerators, academics and policy makers have continued to struggle with this lack of data, or clarity about what an accelerator is. This has been antagonised by the speed at which accelerators have developed and changed, often leaving research and policy behind.

However, what remains a constant throughout the discussion of accelerators is that they are programs intended to accelerate the development of early stage companies or ideas. Generally speaking, an accelerator is a fixed term program that usually lasts from three to twelve months. It provides a combination of education, mentoring, and networking, often with investment. It is distinct from other forms of investment and incubation, such as angel investing, grants, or incubators.

Despite accelerators still being a relatively new and rapidly changing phenomena, they do have characteristics in common, and the research into accelerators, mine included, has tried to identify these common features.

Firstly, accelerators are not incubators. In a nutshell, an incubator is a building that offers subsidised rent, business support services, and other benefits to early stage businesses. It’s main source of revenue is rental from its tenants. Whereas an incubator is a building, and an accelerator is a program, both are forms of incubation, being methods to support early stage businesses.

For something to qualify as an accelerator it needs to have a number of characteristics:

1. A fixed term program, with a beginning and an end

2. A cohort of startups or participants

3. A diverse group of mentors to support the startups

4. Mentoring to transfer tacit knowledge

5. An education program to transfer acquired knowledge

6. A selection process so that the cohort are perceived as the best in class.

It would be fair to say that if something does not have these features it is probably not an accelerator. Equally, I am in favour of these criteria being used to define ‘accelerator-like programs,’ which may not be full accelerators but can still achieve some of their outcomes, using these approaches.

The reasons these criteria are important are that accelerators create efficiencies for everyone involved, and that is a big reason they are seen as useful.

What is a startup?

A discussion about accelerators also involves discussing startups; the cohorts in accelerators are usually made up of startups, and accelerators are often referred to as ‘startup accelerators.’

The term startup is widely used in the tech and innovation sector to refer to early stage companies. While SMEs are defined primarily by size, startups are defined by aspiration and approach.

The European Commission definition of an SME is based on headcount and balance sheet, with a Small Enterprise being a company with fewer than 50 employees, and a turnover of less than 10m Euros, and a Medium Sized enterprise being a company with fewer that 250 employees, and a turnover of less than 50m Euros. They observe that SMEs represent 99% of companies in the European Union.[1]

Forbes[2] addressed the question of what a startup is, reflecting that it is partly a state of mind. They suggest it is not about the specific age of the company, and could be 5 years old but probably not 10 years old, according to Paul Graham, founder of Y Combinator. It is not about the size either, because startups can be worth hundreds of millions of dollars. It is about the expectation when the startup is founded that it will do something different, be disruptive, and quickly grow large. This is the main differentiator from an SME.

In the context of this discussion about accelerators, ‘startup’ will be used as a shorthand for any company or entity in the cohort of an accelerator program. As accelerators have developed and evolved, not all the companies or individuals in an accelerator program will in fact be a startup as defined above. Some may be individuals with early stage ideas, others may be academics, or more developed companies. What they are likely to have in common still is an aspiration to grow quickly and to be accelerated, rather than to develop organically over a longer period of time.

What are Incubation and Incubators?

The casual and inaccurate use of the terms incubator, incubation, and accelerator, can cause confusion when discussing accelerators, as they are often used interchangeably by those who do not understand the distinction between them. However, the distinctions are clear.

Incubation is the process of supporting an early stage company to improve its chances of surviving into a more developed business.

An Incubator is a building that rents space to companies along with some degree of support, often consisting of subsidised rent or business advice.

An accelerator is a specific approach to providing incubation as a process, and can be run within an incubator.

However, an accelerator is not an incubator: an accelerator is a program, and an incubator is a building. An accelerator can therefore run within an incubator, but not vice versa.

Accelerators and efficiency

Accelerators, when they’re done properly, convene different groups of stakeholders around their program, and facilitate engagements between them in a very efficient way. Firstly, by running an application and selection process they convene a cohort of people or startups that is the best in class in that context. The selection process therefore needs to be the result of a wide and open application process, and an open evaluation by respected individuals. This matters because the perceived high quality of the cohort becomes a key point of value within the accelerator, forming the social capital that is used to help the accelerator attract mentors, investors, and others.

Another key aspect of accelerators is mentors. Mentors are an important part of accelerators because they bring tacit knowledge and weak tie networks to the cohort. So whilst the cohort is of value to the mentor network, the mentors are of value to the cohort. Without both, neither would find much value in the accelerator.

Mentor networks need to be very diverse because they need to bring non-redundant information to the startups in the cohort. Non-redundant information is knowledge the startups don’t already have, but also that they won’t receive again from other mentors. If all the mentors work at one corporation, they are likely to have strongly overlapping networks and knowledge, which will be less valuable to the startups. If they are a diverse network of people with few overlaps in their networks or knowledge, they will bring a lot more non-redundant information to the program, and therefore more value.

Consequently, it seems that the real value in an accelerator program lies in its ability to gather together sources of non-redundant information, which is valuable to the various people involved. As that value lies in the social network of the accelerator, it is social capital. That social capital can then deployed to reward people, such as mentors, by giving them early access to the novel and non-redundant information embedded within the cohort. And it gives value to the startups by offering the same in the mentor network.

The mentors also bring value to each other, being a diverse group of interesting people not already connected to each other. So accelerators need to understand that in return for sharing their information and networks, mentors need to be rewarded with a number of benefits:

1. Access to new non-redundant information (the startups)
2. New network ties (the other mentors, sponsors, etc)
3. Social validation (promoting the mentors on a website, championing them at events)

Failure to ensure the mentors are appropriately rewarded for giving their time and knowledge to the startups will result in mentors losing interest quickly, or failing to engage at all.

The way accelerators gather startups together and then manage engagements with them rapidly during a short fixed-term program creates the same efficiency as a university gathering students into classes. Mentors can address all the startups at once, so knowledge is transferred efficiently. Also investors, corporates, and others can meet the whole cohort in one go, rather than 10 startups arranging to meet each individual separately. The saving in terms of emails to fix meetings, and the actual meetings is considerable.

If an accelerator is also backed by a corporate, or supported by a government agency, then you can see how it is able to connect an ecosystem together quickly and efficiently. All of this can happen without ever investing in a startup, or looking at investment returns as a measure of success. This is why I talk about accelerator-like activities, and why measuring the outcomes of accelerators needs to reflect more closely the reason it was set up in the first place.

What do accelerators do?

Two types of knowledge are delivered by accelerators. Mentors mainly transfer their tacit knowledge — what they have learned along the way, over the years, which they condense into rapid fire feedback in mentoring sessions. Workshops, training sessions, and other more structured education transfers acquired knowledge, which is codified information that can be taught. Accelerators are again defined by offering both tacit and acquired knowledge, through this combination of mentors and structured education.

It does all this in a way that creates efficiencies for those delivering and receiving the value it creates by forming a cohort of participants. This cohort functions like a class in a university, allowing one lesson to be delivered to a group of startups at once, rather than individually multiple times. It also creates a focus of attention for the other stakeholders that form an ecosystem around the accelerator, offering them the opportunity to meet a validated group of startups at once, rather than having to find and meet them all individually.

In this respect, the accelerator performs a function for the wider ecosystem in which it exists by selecting the best startups from a larger group of applicants, and convening them in one space so that investors, corporates, and others can meet them. It also selects and convenes a group of mentors, who provide advice, knowledge, and new contacts to the startups to help them develop.

Therefore, a defining feature of accelerators is this use of mentors to support the cohort. Mentors are typically not paid, and should provide a very diverse network, with a broad range of knowledge and experience in order to address the lack of knowledge and networks of the early stage startups.

Accelerators are typically funded by investors, corporations, or government agencies, in order to support innovation, source investments, or identify new innovations. They create returns in the form of economic development, investment returns, or disruptive innovation and new technologies.

Despite the considerable increase in types of accelerators, and other organisations that incorrectly identify themselves as accelerators, a simple generic answer to the question of what an accelerator is would be that an accelerator is typically a 3 month program with a cohort of 8–10 startups, supported by investors or corporates, and with a large, diverse group of mentors.

The benefits of ‘accelerator-like’ programs

This quick overview shows how an accelerator can be used to convene people and manage engagements efficiently. It can also be deployed to communicate knowledge in the same way as a university gathers classes of students together for lectures.

Accelerators can also provide a mechanism to survey and filter a large number of innovators, whether they be startups, academics, or individuals. By running an open competition for prize that is perceived as valuable, if it is properly publicised, the accelerator will see a large number of applications which is equivalent to a survey of activity. By selecting the best of those, it creates a validated cohort which becomes of value to others, such as mentors and investors.

When a corporation wants to embark on corporate venture, the question arises of how to find deals, how to funnel and filter them, and then how to engage with them. This type of accelerator-type approach offers clear benefits in terms of efficiency. However, it would be a mistake to confuse that approach with the original investor model of an accelerator. Investing in the startups for a return may be far less valuable than using the process to find innovators and startups to work with, or to invest in much later.

When accelerators invest systematically in their cohort, they also by definition reject everyone else who applied. This works for some models, but if the aim is to build an ecosystem of innovators then rejecting most of them can defeat the whole point, so investing can be a distraction from far bigger outcomes.

Accelerators are a good idea if they are designed to achieve very specific outcomes that match the needs of those setting them up. If they are cut and paste efforts then they can easily fail to achieve what was initially intended.

But the basic idea of clustering innovators, building large weak tie networks of mentors, and mirroring the efficiencies of a university are of great value when properly understood and intelligently implemented.