For the past two decades, accelerators and incubators have held out the promise of supercharged growth for startup founders and corporate innovators alike.

But does the model behind these organizations still work? And are physical facilities even relevant in a post-pandemic world of hybrid work and rapid advances in AI? We explored this subject at the first Global Ecosystem Summit in Fall of 2024, and generated a broad range of ideas for rethinking and reimagining these important elements of innovation infrastructure.

In search of answers, two NorthGuide experts – Vice-Chair Terry S. Stuart and Strategic Advisor Dan Herman – sat down for a conversation recently. Dan and Terry each have extensive experience with accelerators and incubators, and unique insights into the challenges and opportunities they present to innovators today.

Here’s an edited transcript of their chat.

Terry – Dan, you’ve been around this space for decades. Tell us a bit about the history of incubators and accelerators.

Dan – The first true startup accelerator was Y Combinator, first in Massachusetts in 2005 and then in Silicon Valley. YC became the poster child for formal, cohort-based startup assistance.  If you were an entrepreneur, usually a young entrepreneur, you would apply to YC and, if you were lucky enough to get in, you would get access to structured programming, support and education, and access to a network of mentors, investors and customers.

From there, the industry really flourished. Over the course of a decade, it went from one organization to an industry where, if you look around the world today, there are thousands of organizations who purport to be startup accelerators and incubators. It’s why some have called this the “incubator-accelerator industry complex.” It’s taken on a life of its own.

And in so doing it’s gone from something that was quite unique and targeted to a really select group of entrepreneurs, to all of a sudden becoming open access. Every sector, every geography, everyone… and in so doing we no longer have a discrete model of what an accelerator is but rather dozens if not hundreds of models to address each one of these populations.

We democratized the accelerator industry and gave everyone access. That might look good, but whether it achieves a result should be discussed. Because ultimately our economy desperately needs a growth engine and it should have high-potential founders. So we need a way to support the companies and founders who have real tangible growth potential. AI will likely help us identify those founders, but then we’ll need to concentrate resources - in particular high-quality mentors and programming - towards them to maximize our chances of winning a global economic race.

Terry – Okay, but is democratizing the accelerator industry a bad thing? Do these organizations have an impact?

Dan – So, that’s the biggest question going - do they work? And by extension, what impacts do they create?

To answer that we need to agree on what the point of these organizations is. That should be easy but, let’s be honest, it isn't. That’s because the democratization we’ve seen in this space is tied to the fact that the model has morphed from its origins as a privately-funded activity, to one that, in many places, is funded significantly from public ,i.e., government, sources. The result is that the return-related aims that motivate rigour amongst private funders are now joined by more ambiguously measured returns related to community and regional development and the seeding of entrepreneurial culture in those places by public funders.

The co-mingling of these two unique sets of aims means that the data we look at with respect to impacts is difficult to parse. Back in 2014-15, I led a research program looking at Canada’s then 100+ incubators and accelerators to see if we could answer this question. Unfortunately at the time the data being collected was not sufficient for us to come to any clear conclusion.

Fast-forward though, and there’s really fascinating research that came out of The Wharton School last year. Two professors looked at 8,500 companies from around the world who had gone through startup accelerators and compared them to ones who hadn’t. And that research does show a positive impact on the amount of funding that participating companies are able to attract, the revenue growth they experience post-graduation, and employee counts.

The problem is, the difference is quite small – about a three per cent increase in venture funding raised. And that doesn’t quite fall in the margin of error, but it can be skewed very quickly by a few successful outcomes.

Here in Canada, ISED (Innovation, Science and Economic Development) sponsored a study, also in 2023-24, that found relatively similar outcomes. They found that companies who participated in accelerators attracted a small, positive difference in funding raised. If memory serves, they measured a 13 per cent increase in revenue in the first year post-participation.

Both of those studies show that there can be positive impacts on direct economic returns. But they also leave lots of space for questions as to which types of companies see those impacts, and which type or structure of programming is most effective in creating it. And evidently this doesn’t capture the community and cultural benefits that these spaces can catalyze.

That said, if you really want to be objective, look at our tech ecosystem here in Canada and ask which one of our leading tech companies has gone through one of these accelerators. Shopify didn’t, Real Matters didn’t, Lightspeed POS didn’t, Cohere didn’t.

This is where a lot of the controversy comes from: do these programs actually lead to success?

Now we have some data that says, yes, they do, but even here all the authors of these studies go to great lengths to add the caveat that success depends on program design, the fit of a company or startup stage, their sector, their tech staff, and the support that they get access to. Nuance matters, in particular as it relates to policy makers who are seeking to support entrepreneurship at the community level.

Terry - Okay, but can we agree that the model of the startup accelerator likely needs to change in this era of AI?

Dan - Sure - whether it’s AI today or COVID before, the model needs to evolve to ensure that scarce resources are allocated most effectively to get the results we want. So COVID certainly changed things geographically, in terms of how we organize ourselves, where we work and for accelerators, how important co-location is. Similarly, I think the current infusion of AI suddenly changes the speed at which we need to operate, and how selection processes and advice gets delivered.

But in my mind, none of this changes the fact that the best accelerators in the world still have a value proposition. I’m stealing this from Ben Horowitz, but the best way to learn how to build a startup is to do it yourself, and the second-best way is to spend a lot of time with the people who’ve already done it. At the end of the day, that’s one of the secrets of really successful accelerators: They attract really great mentors, really experienced founders and investors to work with startup founders.

One of the most successful accelerators in Canada is the Creative Destruction Lab at Rotman at the University of Toronto. The beauty of their model is both how tough they are on founders and the calibre of mentors they surround them with. You only get to the next level if you’re able to pass this gated approach where experts and critical thinkers in your field are evaluating you. This is where the wheat and the chaff start to separate. But ultimately you can’t build this model everywhere as there’s only so much of this talent - both top founders and top mentors - available.

Terry – That’s fair but are we, and should we be constrained by geography anymore? And why don’t we see more incubators and accelerators adopt a notion of global connectedness to get our founders integrated into global networks?

Dan – I’m in violent agreement with you, Terry. The global piece is the next step because a) it’s clear that the rest of the world has caught up, and b) we desperately need to diversify our customer markets.

It used to be that the planes would go from everywhere else to Silicon Valley. There were a few exceptions, for sure if you were interested in life sciences you’d go to Boston or San Diego, cybersecurity, you’d go to Tel Aviv, and if you were interested in e-commerce, you might go to Singapore.

But for all intents and purposes, the locus was Silicon Valley. But as ecosystems have flourished on every continent, and as we got a bit more realistic about the competitive dynamic you face when you focus solely on one market, we’ve  learned - or we should be learning -  that you can’t just be there.

We need to build global connections to the big markets of the future, which are in Africa, Southeast Asia and Latin America. We need to change our perception of these emerging and frontier markets and realize that talent, customers and even competition is just as likely, maybe more, to come from these places going forward. We need to get on a damn plane and see what’s going on in Bogotà, Jakarta, Kigali or wherever.

Terry – As we wind up this conversation, Dan, what are the key things readers should be thinking about when it comes to the future of incubators and accelerators?

Dan – I’ll give you three.

The first is that there’s a flight to quality under way - the best founders will go to the best accelerators, regardless of geography. And the best accelerators should be looking for the best startups regardless of geography.

The second is internationalization and the need for our founders to get more plugged into global networks, in particular in emerging markets. This means our accelerators need to get more plugged into their global peers in those same markets. We need to build the connective tissue between our entrepreneurship and innovation systems as a first step towards trade diversification.

And the third is about experimentation or disruption. We spend a lot of energy and resources supporting companies that solve direct B2C or B2B pain points. This is necessary. But given the challenges we have in society, notably with housing and food insecurity, how might we take the brains building ‘regular’ tech solutions and get them to focus on these more existential challenges.

Often we’re told that you need to be an industry expert to build a properly targeted solution, but two of the most disruptive companies of the post-2000 era are a mobility company (Uber) that was built not by taxi drivers, but by some folks who thought there was a better way to do taxi service, and what is ultimately a hotel company (Airbnb) that was not built by hoteliers.

So how can we apply this notion of disruption to other challenges? What’s the solution to housing? Is the solution to housing going to come from developers? Maybe in part, but it’s just as likely to come from entrepreneurs who understand how to scale processes, break down complex processes and tap into the sociology and psychology that drives us.