Want to understand start-up failure? Look at airplane crashes
“It’s hard to believe. It looks like an airplane crash”.
Of the many insights from that day, this is the one I don’t forget. As all co-founders meet to decide about the future of the company, machines are still rumbling on the work floor. Production team — remaining 5 employees after an initial 50% workforce cut — are still hopeful of a turnaround. But the decision we had just made would make those the last moments of our frozen food start-up.
After weeks of consecutive negative outcomes — we had been denied another round of working capital loan, sales contracts stalled, and investors lacked confidence on additional funding-, we came to the conclusion that we couldn’t do it anymore. We should stop, lay off the remaining workforce, and liquidate the company. We had just celebrated our first year and entered then to the statistics of start-ups dying young.
At that moment, the airplane crash metaphor was just a form of transforming facts into an image in my mind. Just like a regular flight, we had planned our moves and resources ahead of time. We knew we would face turbulent times, but were confident about our product, our strategy, and our market fit. From our experience with other start-up projects, we knew most of what could go wrong, where the pitfalls resided and, more importantly, how to avoid or overcome them should they arise.
For the next couple of months, I would routinely sit down and reflect on what had just happened. What went wrong? Why didn’t we succeed in this venture? Why didn’t we achieve what we planned? I wanted to approach this case as someone that peels an onion, going to the core of everything. While doing so, I felt I needed tools to help me, some methodology that would guide me to the core(s) of the failure we had just been through.
The best tools I found were the ones related to incident assessment, mostly used in fields such as aerospace, health, and corporate safety. They helped me find out the root causes for our failure, when and how they appeared, where they stemmed from, and so forth.
Introducing the Swiss Cheese Model
This was when I came across the Swiss Cheese Model of Accidents, a simple yet interesting tool used to assess the occurrence of incidents in high complexity systems, such as Air Traffic Management (ATM). Interestingly, it turns out that the understanding of the dynamics of airplane crashes are just different from what I had in mind. The outdated picture of an accident as a byproduct of a sequence of faulty events is just not accurate. It had been like this a long time ago, but relatively new tools are currently in place to better grasp how these incidents happen, and the Swiss Cheese Model is one of them.
Although the model is initially intended for the analysis of accidents, extrapolating such understanding to company failure is a smooth ride. Instead of an airplane crash, we would have a company crash. Instead of safety mechanisms to prevent an air accident, for example, we would have them to prevent business failure. For the sake of an easier understanding, let’s break the definition above into parts:
“Well-defended system”
In this model, companies should have defense mechanisms in place — and keep them optimal — , which would safeguard them from incidents. In our parallel with start-ups, these defense mechanisms would work as barriers to prevent general business failure. These defenses are represented by vertical plans, which work as barriers to avoid hazards become losses. However, these plans are not perfect, which brings us to the next point:
“Failure”
In this context, failures are understood as “latent conditions” and “active errors”. Such conditions and errors would be the holes in the vertical planes, creating a similar image of a slice of swiss cheese, hence the model name. Latent conditions “are present in the organization long before a specific incident is triggered. Most of them are a product of the organization itself, as a result of its design or as a result of managerial decisions”. Active errors (or failures) are considered short-term breaches and may be created by the errors and violations of front-line operators. According to the model, these holes vary in size and are not static, roaming around the same plane and repositioning themselves until becoming a breach of that plan by an oncoming threat.
“Conjunction of several contributing factors”
When holes in these many plans align, allowing hazards to become losses, we have a major breakdown in the system. It is not until many of these plans are crossed by the same beam of threat that the system fails. As a normal dynamics, such beams are constantly coming against these plans, but for the misalignment among the several holes in the many plans, the system acts in order to push the threat away. When one or more defense mechanisms fail, there are still several layers of other types of defenses to prevent a threat from becoming a loss.
“Different levels of the system”
This is an important remark to understand that the weak spots in a chain of failures may be at different levels. An operational error — thus an active error — may well trigger the Armageddon scenario that started with a latent condition, or even an active error in a higher level, such as the strategic one.
You’re not alone
While trying to understand my own last failure, I started being curious about start-ups failures in general, especially to what extent founders and co-founders really understood what they had been through when their business didn’t succeed. In short: “Why do start-ups fail?”. Or even better, do start-uppers really know why they fail? Do they really dig deep enough into the several causes and consequences of decisions made in the course of their start-up existence? Do founders and co-founders really learn as much as the experience makes available?
The search for data about start-up failure
Apart from “top-something” lists made by someone(s), I could not find on the internet any reliable data-set about start-ups life and death details. It turns out it is not easy to find data about start-ups failure. When available, they are not standardized, nor structured, and mostly hard to dig. The largest database I found was from CB Insights, that put together around 100 post-mortem statements from ex-founders and ex-co-founders of failed start-ups. From this sample, they built a Top 20 reasons why start-ups fail. It is clear that it is not an easy task to build up such statistics from non-standardised data, but many of the reasons for failure presented in their studies look more like consequences, not causes. On one example, “ran out of cash” seems to be the second most frequent reasons presented by founders for their start-up to fail.
But running out of cash is a side effect of (many) other things that went wrong, not a cause. It’s just like saying that the patient died because he/she stopped breathing.
You may have run out of cash because you failed to make new loans or to receive more investments, which in turn may have been because your sales stalled and you didn’t achieve the sales volume milestones that would have triggered another round of investments or loans. And your sales may have stalled for many reasons, from fierce market competition to your overestimation of market acceptance of your product in the price range you needed to keep up sustainable margins. The list of reasons could go over and over, but it would demand time and efforts from former founders and co-founders to actually come to it.
The level of details in the information provided by the post-mortems definitely does not allow us to reach the core of each case, mainly because even the founders and co-founders may have not done this “onion peeling” exercise. However, it is possible to realize that the errors that led to the failure reside in one or more of these domains, not necessarily in this order:
- Founders’ experience (or lack thereof)
- Product/service design
- Teams
- Business Strategy
- Production process
- Business Plan
Some of these topics may seem vague and too open (i.e. “business strategy”), while some others may look like they are a subset of another topic (i.e. production process and business strategy). This is roughly, again, because of the lack of details from the information provided. But it is nonetheless valuable and important to sketch up what would be the “cheese slices” should we apply the Swiss Cheese Model to the entrepreneurial world.
But once we have a glimpse of what the cheese slices would be in the Entrepreneurial Swiss Cheese Model, it is also interesting what the breaches would be. For the lack of detailed information already mentioned, I share with you some of my own findings from our most recent failure. The list is not extensive but I cut it short for the sake of keeping the article short. I keep the “latent conditions” and “active errors” framework to illustrate the use of the model.
Latent conditions:
- Attachment to product and to product history — someone who has created a product or service from scratch knows how difficult it is not to have a biased view over it. Founders may be excessively attached to their products, having a hard time to perceive market reactions and inputs;
- Excessive optimism — while optimism is natural to entrepreneurs, like everything in life, entrepreneurs should keep an eye out for a balanced amount of it. It may blind them from seeing reality, just like product attachment;
- Lack of entrepreneurial drive — many entrepreneurs fail to shift from an employee mindset to an entrepreneur one. Self-motivation is key to an entrepreneurial spirit; grit and drive are raw material to endure the hard path of entrepreneurship. Curiosity and creativity fuel the engines of any entrepreneur and all of this builds up to the entrepreneurial drive. Lacking one or more of these skills may detour you from success.
- “Not seeing the business as a priority” — starting up your venture as a side job may be the most common scenario in the entrepreneurial world. I don’t have any data about it at hand, but I would guess so. While a traditional 9–5 job may give you the stability and financial security one may need, there may come a time in which one should devote him/herself to the venture path, making it a priority, risking it failing badly.
- (Mis)Alignment amongst founders: this one is broken into 2 sub-items:
- first, although may seem obvious, is that the skills from founders should match (near to) 100% with what the company needs. Drive, passion, grit, and curiosity make up the basic founder skill set. Ask yourselves as founders if the timing is right for each of you to join the company. Or if all of you should play as founders. Maybe some of you are better off as investors offering smart money, with key strategic skills, but without daily management. But once co-founders join each other, their skill set should complement each other and be useful to the moment of the company.
- second, expectations (financial or not) should be leveled. Uneven levels of expectation bring uneven levels of stress when things start to run different from the plan and this may become a snowball, bringing more stress to an already chaotic environment.
Active errors:
- Stressing the financials: start-ups are by nature extremely unpredictable when it comes to resources. While it is common to detour on your initial (and constantly amended) financial plans, thresholds should be respected. This active error may walk hand in hand with the “excessive optimism” latent condition, and letting them align is a strong breach of business continuity.
- Making decisions on flawed market perception: starting to play in a different field demands many hours of study and deep understanding of how that field works. What are the specifics about consumer habits; who are the strongest players; what are the go-to-market strategies; who else started in this market recently; who failed, and who succeeded? All these questions are worth finding answers before or during your venture journey; In not doing so, your actions as an entrepreneur will cause harm to your venture journey;
- Ignoring business complexity: it doesn’t matter if your start-up develops new sensors for sky-rocket burners or manufactures dog cookies, all business has its own deep level of complexity. If it is not on the production, it may be at the distribution; should it not be at distribution, it may be at marketing; not marketing, then maybe after-sales; but never take for granted the complexity of any business, despite how simple their core activity may be;
- Hiring low-qualified of workforce: start-ups live by its nature on a tight budget and may not be able to afford to hire a highly skilled workforce. While much of the activities may be handled by founders and co-founders, there will be a time in which you may need to delegate (part of) what you do, especially when scaling up. Although there may not be many other solutions, reaching out to lower qualified teams — hoping they would learn from you — may be a tricky move, given the fragile nature of the start-up business.
As you may see, active errors derive constantly from a misperception of what your venture’s latent conditions are. With such a blurry view, your actions to try to fix such latent conditions may cause you to make bad decisions, which will, in turn, become active errors, in an endless cycle of trying-to-fix-causing-more-damage.
Finally, it is important to think of the Swiss Cheese Model as a tool to help you build a healthy start-up, not only to work on hindsight. It is used to assess incidents, which are by nature a consequence of when the defense mechanisms failed and there’s not much else to do. But there is much value when you understand the model and use it while running your business. Consider what would be your planes, how strong they are — it is ok to have fragile ones in the beginning, given the nature of start-up businesses — and how to maintain them. Learn to identify what latent conditions your start-up may have and work towards them. Learn how to foresee active errors and avoid them. But never, ever, give up on trying. And never, ever, let the holes in the planes align.
