Antifragility in Software and Websites

In his eponymous book, Nassim N. Taleb defines "antifragility" as the property of systems to not only survive stress, uncertainty, and chaos but to actually benefit from them and become stronger. What could this look like in our digital lives?

What is Antifragility?

Companies, political systems, and we as individuals are exposed to various stressors and pressures. While many risks are predictable, some—especially the very large ones—are not. They catch us unprepared. This can happen for various reasons. Sometimes, events have never occurred in this form before. Like a black swan, they appear where only white swans existed before.

But how do we deal with such events when they are unpredictable? Especially when they have the greatest potential to upend our lives, our businesses, and our society? How should we organize systems so that they do not collapse under pressure and unforeseen events?

This question is also posed by Nassim Taleb in his "Incerto" series. He notes that intuitively, most people think that the opposite of "fragile" is "stable." Like on autopilot, they try to construct systems to be as stable as possible, forgetting two points. First: stable systems are only stable up to a certain point. Second: a stable system ideally does not change at all. But this also means: it does not benefit from the unforeseen.

For the actual opposite of "fragile," we previously had no word. A system that does not break under pressure but actually benefits from it is therefore called "antifragile" by Taleb. He finds it, for example, everywhere in nature. When we train, our bodies gain muscle mass, but not just enough to perform the same exercise better later—rather, it goes beyond that! Trees strengthen their trunks at the points where they are most vulnerable in a storm—but with more than the strongest storm would have required. Trees thus benefit—up to a certain point—from events that are worse than any previous ones. Because while a storm-tested tree has reinforced the important spots in time, other trees fall over as soon as they are newly positioned in an exposed location.

From these examples, Taleb draws lessons for all kinds of systems—from personal fitness to financial markets to forms of government and workplaces. In this article, we want to shed light on the digital world—specifically websites and software. How can we make these areas more antifragile so that we not only remain stable during the next black swans but actually benefit from them?

The Least Common Denominator

Perhaps some of you remember the time of Flash websites in the late 2010s? Flash was a trendy new technology that allowed for the creation of websites that would have been unthinkable with other available technologies like HTML and CSS: animations, transitions, and interactive elements provided an unprecedented user experience. And quickly, anyone who did not incorporate at least some Flash on their website was considered outdated.

But almost as quickly as the phenomenon appeared, it was gone again. Flash was a proprietary technology from Adobe that made all website operators dependent on this one provider. While it was supported in many browsers, cross-platform, and provided a refreshing change at a time when developers struggled with the various requirements of different browsers (yes, we're looking at you, Internet Explorer!), support gradually waned as more security vulnerabilities emerged in the technology. Additionally, new devices came onto the market, particularly the early iPhones, which did not render Flash content well. Thus, it was consistent that Adobe pulled the plug in 2017 and phased out Flash in 2020.

This meant that many operators had to redesign their websites almost completely. And what technologies did they use for this? Primarily HTML and CSS, both of which have existed practically since the dawn of the internet and have undergone many phases of backward compatibility.

Some were smarter and relied from the start on resilient technologies like HTML, CSS, and JavaScript, using Flash only where it was not mission-critical. Flash had its advantages—but building a complete web presence on a new, not globally supported technology was costly and educational for many.

What is the Flash of today? Well, from a usability perspective, there are always new technologies that promise to shake up the website and software market. But they are often associated with unnecessary dependencies or limited support. JavaScript, for example, is a robust technology for animations and similar tasks, but building entire websites dynamically as "Single Page Applications (SPA)" with React, Vue, or Angular does not seem sensible when considering that at least 10% of browsers do not support JavaScript, whether due to ad blockers, accessibility issues, or user choice. The same applies, albeit for different reasons, to website builders or proprietary offerings from individual software vendors.

It is not a bad thing to build websites and software on the least common denominator. On technologies that may not be "cutting edge," but have been robust, reliable, and widely used for decades. Those who do not want to completely rebuild their digital solutions every two years or exclude important customer groups rely on proven standards. They use new technologies only where no unhealthy dependencies arise. This way, they are prepared for the future—regardless of what new gadgets and screens humanity invents; the likelihood is very high that they will not completely overturn the entire previous development of programming and design languages. And those who have a resilient website that is flexible enough to adapt to new technologies save real money and also consider the second point: room for trial and error.

Trial and Error

Consider this: what groundbreaking invention comes to mind that was based solely on theoretical groundwork and required neither an element of chance nor at least some practical tinkering?

Trying out new things and accepting small losses to realize potential large but previously undiscovered advantages is antifragile action in its purest form. Success is often not predictable. Most breakthroughs arise from small, seemingly insignificant changes. Here, communication with customers and the ability to effectively gather constructive feedback play a significant role. Not least, this is a huge advantage of online communities.

This freedom to trial and error must be affordable. And one needs structures that allow for tinkering and experimenting with what might work. In relation to the software and internet world, this means: Software solutions and websites must be flexible enough to incorporate new functions and elements without jeopardizing their functionality and usefulness. Often, even small changes can make a big difference. For example, we repeatedly experience that the click-through rate (the proportion of users who click on a specific link or ad relative to the total number of users who have seen that ad or link) can change massively with even the smallest adjustments. A single point in the right place can play a decisive role.

Is your website, or is the software you use, structured so that you can quickly and easily add new functions or adjust existing ones? Or are you trapped in a "stack" (a combination of technologies, frameworks, and tools used together to develop and deploy an application) that is so fragile and fixed that every change is a huge effort? How do you plan to benefit from unforeseen events?

Many Cards in Play

In German, there is a saying: "Putting all your eggs in one basket" refers to someone who completely depends on one thing at a certain point. He (or she) can only hope that this card will play out.

This is fragile action. Antifragile action, on the other hand, relies on redundancies and alternative plans. It takes nature as a model, which—so far as efficiency still allows—holds everything important in multiple copies. The creator had a reason for giving us two eyes, two lungs, two kidneys, two legs, two arms, two ears, etc.

In the digital world, extreme risks lie in putting all your eggs in one basket. However, this can be avoided through measures such as:

• Using technologies and software that work on multiple operating systems,
• Using technologies and software for which there are not just a few experts,
• Building multiple independent channels for sales and customer outreach,
• Modular software and website construction with interchangeable components (no unnecessary "break-all-plugins" or "locked-in" providers),
• Using redundant server systems with load balancing and failovers,
• Using geo-redundant offerings (e.g., multiple server locations in different places worldwide),
• Emergency plans, including analog alternatives where possible,
• Test environments for updates,
• Creating regular backups, and
• Meaningful documentation of projects for future developers...

The advantage of these redundancies is not only that your systems run more securely, but also that they are likely to continue running even when regional or international competitors face massive problems (e.g., due to a hacker attack on a central system, local events like the failure of a data center, or due to a programmer's mistake.

Old is Gold

One final thought: Be generally cautious of new providers offering you a revolutionary product that breaks with everything that has come before. The likelihood is very high that this product, this technology, will soon no longer exist.

Nassim Taleb has developed a simple heuristic for this, which he recommends with a wink (and without claiming universal validity). He does something he usually warns against: he looks to the past to guess the future. Specifically, he suspects that a technology will exist in the future for about as long as it has already existed to date. Instead of inventing all sorts of fantastic and exotic technologies for a vision of the future, it is better to continue to assume what has proven itself and remove the novel from it. We will likely still be sitting on chairs and eating with forks in a hundred years, but it is probable that many of today's trendy technologies will no longer exist. They will fade away, just like Flash, which flared up briefly and then disappeared.

In digitalization, many are tempted to completely break with the past and instead invent something innovative and new. This is neither wise nor antifragile. Let us keep what has proven itself. And let us use the groundbreaking innovations that our society produces wisely and moderately. For this is not old-fashioned, but reasonable. And it better prepares us for what we cannot yet foresee: unforeseen events.