Technological progress and the legacy trap
(Legacy Series 1/3)
My series on legacy systems will show how the adherence to old tech, processes and skills is the major threat for companies and economies. In the wake of breakthrough innovation such as blockchain or AI, legacy-strapped China is vying to dominate global business and finance. In three essays I will examine what policymakers and managers must do to capture the promise of the digital age.
In the wake of publishing my book Blockchain Babel in 2019 I got the chance to speak with numerous executives and policymakers about capturing the benefits of this new technology. Most of them were eager to find out more, but very little of them argued convincingly they would be adopting it. They unanimously gave the same reason: legacy systems. One CEO told me that implementing blockchain into his organization would be akin to equipping a horse carriage with a rocket engine.
At the same time, I met executives from developing countries who reported their companies were leapfrogging entire technologies such as landline telephony, mainframe computing, and credit cards. And they proudly described how they were institutionalizing the flow of constant change. The Chinese giant Ant Financial, for example, trashes its entire infrastructure every 3-4 years. The government is a catalyst, marshalling unprecedented sums into the adoption of technologies such as cloud computing, artificial intelligence, and blockchain. It is no wonder that on average China’s legacy systems are 7-8 years old, compared to 32 years in the West.
There was another thing that bothered me. I realized I was talking mainly to experts from finance, cybersecurity, and governmental bodies: An economy’s most critical infrastructure is the most prone to legacy and its problems.
The pressing threat of legacy systems
System outages, costly workarounds, and glitches in a customer’s app can be small or major trouble. But legacy does not only plague individual companies. It is not even a matter of a state’s pure economic competitiveness. Sometimes it is a matter of survival. The Roman Empire fell because it clung to an outdated and doomed system: slavery. Cheap and abundant workforce prohibited the deployment of labor-saving technology. It is why Romans fell short of scaling breakthrough innovation such as windmills and water wheels, both of which were available to them. To be sure, there are multiple triggers for the Roman Empire’s demise, but all started with the economic weakening. It left the empire vulnerable to attacks from the in- and outside.
The same is true for the American Civil War – the south stayed agricultural whereas everything north the Mason-Dixon line picked up steam in the industrialized age. Factories, railroads, banks and other modern institutions flourished as Benjamin T. Arrington describes in detail.
So, what can companies and governments do to remain agile and resilient in the long run given the accelerating technological progress? It is a decisive question that each manager needs to address. This is why I have decided to write this series on what exactly legacy systems are, how they manifest themselves, and most importantly, how they can be managed and overcome.
Today the urgency to deal with obsolete infrastructure is unprecedented. What makes acting now exigent is the unfolding of no less than three carrier technologies and their impending combination: cloud computing, blockchain, and artificial intelligence (AI). The cloud decouples IT-resources from the end-user. This causes an increased need for secure connections and transactions, which is guaranteed by the blockchain. Blockchains keep not only secure, but permanent ledgers, just perfect for building the troves of data needed to feed AI algorithms.
Legacy’s Inexorable Conquest
In principle, we all agree that new systems are more powerful, more efficient, and more secure. After all, nobody of us is using faxes or typewriters anymore, so why are old systems in the backend proliferating at such an unseen speed?
At the beginning of all legacy, even non-technical one, stands innovation. Science unearths natural phenomena; technology harnesses them. And an economy arises from the sum of its technologies and their interplay. While individual technologies have been studied extensively, its principles have at times been neglected. But to understand and fight legacy systems, we need to understand the general mechanisms driving technological progress.
The crucial mechanisms enabling legacy lie in the very nature of technological progress: Despite its recursiveness, meaning new technologies build on existing ones, for companies there is no benefit of having old systems in place. And due to the combinatorial nature of progress, the pool to combine from grows every day. Hence, innovation and legacy-building are accelerating every day. The consequences of that are best described by Schumpeter’s concept of creative destruction, whereby formerly successful tools, processes, and products become a major threat to a firm’s competitiveness.
Yet there are also psychological biases that allow the old to flourish. We live in a world where the old is revered just for being old. A company might be branded as “established in 1910” – never mind it is smaller and less customer-oriented than its 5-year-old competitor. Add to this a dangerous cocktail of scientifically proven mechanisms such as risk-aversion, cognitive dissonance-evasion, or selective perception, and you will understand the power of mental barriers against switching systems.
One particularly troubling bias is the sunk cost fallacy. It makes managers throw good money after bad. Stopping to invest in a doomed project would be equal to admitting that it has failed.
We can gauge how hard-wired our thinking is by inspecting the unconscious metaphors that steer our actions. The act of building software is still likened to the process of constructing buildings, which makes absolutely no sense in world where developed parts can be simply copy-pasted and mistakes ironed out without wasting resources.
The rank growth of legacy systems is often also an inescapable result of solid business decisions, or rather their aggregation. Vivid M&A activities, the imperative of backward compatibility, and local, fragmented decisions all lead to rank growth of heritage systems. What is often forgotten – old tech sometimes has specific advantages. COBOL, a fossil among the programming languages, reaches incredible processing speed by its batch-system. But above all, existing systems work reliably. 46% of banks consider any kind of tinkering with the core system a significant operational risk.
Is each old system legacy? And what exactly is “old”?
Is every old system legacy? And how do you recognize legacy in people or customers? Heritage-systems are so hard to define because they are pervasive. According to a Congressional Report from 2016 the U.S. Department of Defence uses 8-inch floppy disks to coordinate the nation’s nuclear forces. Banks worldwide run on IT systems from the 1950-ies. Legacy systems have existed for thousands of years in machinery, processes, employees, customers, and mindsets.
There is no cut-off date when a system solidifies into legacy structure. A host of factors determines whether something is so-called brownfield architecture. Are there enough programmers? Can it be scaled and plugged into more modern applications? Has the soft- or hardware reached its end of life?
Another crucial indicator is the speed with which complexity grows. Complexity in a company’s systems is the main cost-driver and agility-inhibitor. Dependencies grow exponentially with each new software package, each new feature, and each new workaround. Eventually every new unit costs you more, not less.
Complexity also leads to what is known as legacy lock-in. It is a vicious spiral that makes you more dependent on your systems and your providers every time you think you are taking a step forward.
Major inventions of the past such as ERP or relational databases were supposed to improve complexity, but actually made it worse. If you want to know how exactly these new inventions exacerbated legacy, I suggest you read Software Wasteland by Dave McComb.
Heritage systems are a problem transcending the lines of code. Thinking of legacy solely as an IT-operations topic is outdated. First, legacy can manifest itself on multiple levels, including machinery and people. Second, even when dealing with IT-heritage, the impact goes beyond downtimes and processing speed. Lagging product features and external interfaces inhibit competitiveness and profitability.
In business, complexity is always linearly linked to costs. Obviously, there are also entire industries that live off keeping old systems alive. Most large projects are delayed, cost a multiple of what was budgeted, and in the end many of them are left incomplete altogether. But the real danger are the hidden costs. Two out of three CFOs say they don’t have the time for strategic tasks because they are overburdened by tasks such as assembling data and solving inconsistencies, according to BCG. And the opportunity costs in terms of money and new functionalities are even harder to estimate with the consequences no less dire.
Regardless of the impact on costs and security, the danger of missing out on future markets or competitive advantages tops the threat-list. McKinsey found that executives attest to new IT more potential to bring in new revenues than to fix expenses. The same executives, however, also admit companies don’t act accordingly.
If you want to know how get on top of old systems, stay tuned. In the next edition of The New Frontier I will look at the three major types of legacy and the best strategies to combat them.