Why most accelerators struggle in Africa
Startup accelerators — you know, the 3 months where 10 high growth startups run through an intensive mentorship program that results in a demo day to raise lots of money — work relatively well in most first world markets. Pioneered by Y-Combinator and scaled by Techstars, the model has been applied all over the world but achieved only mixed success beyond the few big names.
In our local South African market, accelerators have been running for almost 8 years but without any notable exits. Accelerator programs typically promise “12 months of growth in 3 months”, but most local cohorts take 12 to 18 months before they actually raise capital — by which time they’ve lost any of the momentum that they would have built in the program.
It’s not for want of trying either. Heck, I have mentored on multiple programs, run product and iteration workshops on others, participated in Techstars NYC Blockchain program and have helped design multiple others with the Lean Iterator process. I’ve probably been directly involved in one way or another with 13 different accelerators, and that’s not counting demo days.
Despite clearly believing in the purpose of accelerators, my experience has led me to conclude that the model doesn’t work well in our market. Here’s why:
The 3 month slice
Accelerators were designed in thriving startup ecosystems. They’re a 3 month slice that is both preceded and followed by a lot of funding and support in their markets. At the first 2005 YC batch there was 20x more funding going into US startups than there is in 2020 across the whole of Africa combined.
Our ecosystem doesn’t have the support before, or competitive funding after the program to sustain the same model. I only know the stats from a few accelerators, but I’d be pretty confident that less than 20% of startups from any accelerator has raised over $250k in the following 6 months after their local demo days. And the startups that do probably would have anyway. The accelerators become islands of activity followed by lots of frustration. With limited investor options, the good investors don’t need a demo day to get access to good startups as they all apply anyway.
High variance and low volume of entrepreneurs
First world markets are very homogeneous compared to emerging markets. Everyone has access to standardised education. Access to computers and internet at home is common, as are disposable income and have access to credit. In emerging markets like Africa those are not a given and the majority of the population, where we hope to find our great entrepreneurs, don’t have access to those things growing up..
When you combine that context with the fact that unemployment is so high (over 55% for youths in SA) many of your top achievers will get stable jobs and hang onto them. Very few actually have the safety net to be able to take the risks that entrepreneurship requires.
While high unemployment could be a driver for entrepreneurship, in reality a scarcity mindset will more likely result in small businesses and not high growth startups.
These factors mean that the typical intake for an accelerator has a huge variance in the experience, maturity, sector, and team size of the startups coming in. (One program had a 3 person idea-stage company in the same cohort as an 11 year old 50 person company…)
Scarcity of early stage funding
Our ecosystem is small and diverse which does not attract large amounts of capital. That which it does attract is risk averse and tends to focus on the later stages. Entrepreneurs in poorer markets typically can’t draw on their own or family funds to get going unless they’re part of a privileged few. On top of that, angel investors are few and far between and often come from corporate backgrounds — they don’t understand startup risk properly, which leads to poor advice and worse deal terms.
Even at the series A stage, most funds need to keep enough aside for future rounds too as there is a good chance that they’ll need to fund later rounds as well. Bottom line: There aren’t enough exits to attract enough funders to create healthy competition.
This is slowly starting to change as international investors are waking up to the opportunities here. But for the most part, startups are caught in a bootstrap desert through their early stages before they can get to the low-risk late-stage investors.
Some are exceptional outliers
Accelerators do deliver value for local startups, but the first world design, structure and process doesn’t translate well. Those that are finding success in our markets are the ones that are breaking the traditional mold.
Corporate accelerators like Startupbootcamp AfriTech whose primary focus on access to becoming a supplier to their corporate funders, are one way. Another is Grindstone who focus on helping later stage companies becoming investment ready. In general, the more successful ones are focusing their startups on revenue and getting to break-even — not raising funding.
So what then?
If accelerators aren’t the holy grail then what is? There is a lot that we can learn from what they lack to help redesign what more effective startup support program might look like for emerging markets. Much of the thinking is rooted in the Lean principle that synchronization is waste, as accelerators are designed around cohorts.
Note that I don’t have the answer. I’m not going to provide you a better alternative as I don’t know for certain what that looks like. But there are some key principals that a better model is likely to embrace:
As part of our iterative conversations with David Campey and Philip Kiracofe (and plenty others) these insights have grown… these are not mine, but ours.
1) Better stage-matching of support and education
Given the variety of entrepreneurs applying it’s unrealistic for program directors to expect that all of the startups are at the same level. At least some level of entrepreneurial content and training needs to be available as part of any program. You don’t want to force everyone through it, but having it available with people to help is critical. (I think that Startup School by YC is trying to correct for this as they’ve internationalized and are getting a much wider range of founder experience.)
2) Different objectives for different stages
Entrepreneurs are desperate for support and programs struggle to find good startups, so they both compromise on fit for the stage of business. Most programs don’t adjust the content or program structure to the stage of the businesses in the cohort.
In my experience it’s best to arrange startups by the type of risk that they have removed through implementation and sales — that is, how much traction they have. (I’ve written on this before in What every startup needs to do.) If you can look through the hype and story and focus on the facts of what has been achieved or delivered, then the next steps become a lot more obvious.
3) Create clear deliverables, not processes
Designing a flexible program with a single path through it is almost impossible. Turn it around to be outcome or deliverable focused: set clear targets for the business’ stage and then offer an array of tools on how to get there. (This is how we structured the pitch decks at each stage gate in the Lean Iterator process.) As you build out the a collection of tools that can be used, startups can then decide which ones work best for them.
4) ‘Pull’ not ‘push’
Batching is inefficient unless the whole cohort needs the same support at the same time. This is hardly ever the case, particularly in markets with high variance in what entrepreneurs need.
Instead of pushing all startups through the same 3 month curriculum, allow each startup to constantly ‘pull’ the support that they need. Structure the program more like Coursera than a high school, and provide guides, coaches or tutors to help them move as fast as they can. One way to achieve this is to have…
5) Tightly coupled funding
In African markets funding is typically stretched pretty thin. Corporate backed accelerators have more funding, but it mostly covers the program costs (for rent, fees and salaries) and only goes indirectly into the startups — almost never as cash.
Lack of funds can be crippling and even more so in a market where raising an early $20k round from Friends, Family and Fools is almost unheard of. (Only a privileged few have access to this which kinda explains why most SA founders are white males…)
Running experiments costs money. Being “lean” doesn’t mean spending less, it means learning faster — and that sometimes requires more money. If access to capital isn’t tightly coupled to a startup program then the program is just foreplay with no follow-through. It’s playing at startups, not building them. (The exception is if the purpose of the accelerator is to get sales through it’s corporate relationships.)
6) Provide some funding up front
Some of the bigger international accelerators provide ~$100k to startups that get accepted. Not only is access to capital critical, but startups need to spend that capital during the program and not only at the demo day afterwards. This also has the added benefit of having the startups pay for the services they need on the program which minimizes waste in program support.
The Lean Iterator hypothesis is that the best way to identify who to invest $10k into is to give a number of startups $1k and see how they spend it. Run that thinking from $100 investments up to $100,000 backing the best performers each time and you’re far more likely to end up with winners. (It also turns out to be much cheaper than due diligence.)
7) Distribute the Demo Days
In the NYC program I took part in, demo month was a huge distraction on the business, we made a pretty deck but landed no funding related to the day. In the typical accelerator you’ll spend month 1 for groundwork, 2 for growth, and month 3 prepping your Demo Day pitch. In markets where there isn’t the pressure to invest in a startup at demo day, month 3 becomes a complete waste of time. It becomes more for the program than the startups.
Demo Days are not the only way to teach startups to pitch. Instead of having a once off big splurge event, why not have them every month or 2 where the startups can pitch when they’re ready? Invite investors for each day depending on who’s pitching. Another alternative would be to spend that month tailoring the startup pitch for the top 3 investors in the world for them and facilitate warm intros. Yet another would be a video pitch with a live Q&A that caters for geographical distribution.
8) Multiple off-ramps
The one thing to note about the accelerator model is that it was designed to accelerate ‘rocket ships’ and build ‘unicorns’. It was designed for companies that want to raise 10s or 100s of millions of dollars to dominate markets. In Africa we need gazelles, not unicorns. The Silicon Valley “swing for the fences” type company is rarely the ideal in our markets as inevitably, no matter how good you are, you’re going to struggle to raise locally due to limited and risk-avers capital markets. (There are plenty of examples.)
A successful program needs to identify and provide multiple off-ramps for startups that are on different trajectories instead of just throwing 90% to the wolves.
Startup accelerators are like a rail gun: great at what they do in certain scenarios, but certainly not a utility tool. We need to learn from what makes them successful in first world markets to better understand what their shortcomings are in emerging markets. Only then can we experiment and rebuild what will work in our environments. Some people are already tweaking the accelerator model to suit their needs — but I’d propose a more wholesale re-imagining would be better at leaving the baggage (like demo days) behind and building a program fit for purpose.
The hypothesis is that by following the principles laid out above you can build an inclusive, supportive framework to really accelerate the volume and quality of local startups. If done right it has the potential to level the playing field and remove the gate of privilege for startup founders and tap into a far greater pool of talent and opportunity across Africa and other emerging markets.
