Recurring news stories of airlines grounded and social networks rendered offline by a few lines of bad code have made the curse of technical debt known far outside the world of IT. While this may be the first time the phrase has entered the popular lexicon, it’s a long-festering problem known to technology chiefs and ignored by most executives, including chief financial officers. The option of ignoring technical debt is over, and it’s high time CFOs understand, invest in, and solve tech debt.
While there are many types of tech debt, a recent academic paper identified 13. But the problem is relatively straightforward and affects every company that builds or modifies software. As companies add features or develop IT systems, they often take shortcuts instead of a better approach that would take more time. The shortcuts can lead to subpar system performance, vulnerabilities, and outages. And that’s how “debt” builds up over time, much like credit card debt.
In addition to crippling breakdowns, tech debt can open the door to hackers, regulatory non-compliance, or attacks by rogue employees. It can degrade the customer experience and website accessibility, lower search engine rankings, and allow licenses to expire unexpectedly. Beleaguered employees waste countless hours on manual workarounds, eventually leading some to resign in frustration. Companies that seek outside investment may watch potential funders walk away because a code review identifies excessive tech debt.
As tech debt builds up, each new release, iteration, or innovation becomes slower to come to market and more expensive. That reduces a company’s ability to keep up with competitors or respond to changing consumer expectations. In acute cases, it becomes like a financial debt spiral, where the only solutions to a compounded problem are extreme and unpalatable for all stakeholders.
There are reasons to fear that tech debt is now piling up at an accelerating rate. As customers demand ever-more features and functionality, vendors naturally try accommodating them. In North America, layoffs and budget reductions are forcing tech teams to cut corners or choose the quickest solution rather than the best and most effective one. After a time, these teams get overwhelmed.
The good news is that tech debt can be eradicated.
Smaller companies or startups can use newer or more modular tech stacks. They can create systems to address tech debt as the corporate culture matures.
Those responsible for maintaining and optimizing a company’s software must be empowered to push back on timelines or budgets likely to create tech debt. Too often, product managers find it difficult to push back on sales teams and executives who demand additional features quickly to win a new contract or hit an internal deadline.
Likewise, companies must create and maintain effective processes for code reviews, non-functional testing, and regular security audits. New tools, including artificial intelligence and automation, can help.
Those responsible for maintaining and optimizing a company’s software need to be empowered to push back on timelines or budgets that are likely to create tech debt.
Some may use structures such as pair programming, where two developers work on the same code simultaneously, constantly checking each other’s work. Here, too, AI is increasingly helpful.
What’s key is that tech debt needs to be considered as a part of product decision-making, backlog-grooming, and sprint planning. The question needs to be asked — if something is done one way, what types of problems are being stored up for later? And what impact does it have on the ability to develop other future products?
Executives and salespeople can misunderstand the tradeoffs between speed and quality. Product managers need to put more effort into quantifying and communicating those tradeoffs. Say a product team can get a feature to market in four or eight weeks. Four weeks might mean a team is 10% slower to deliver subsequent features because of increased accrued tech debt.
Having clear metrics for code quality and test coverage and agreeing in advance to targets and “red lines” can ensure decision-making is more data-driven.
Periodic tech debt reviews should be used to examine product decision-making processes. If too much debt has built up, the future impact of the debt is not being effectively understood.
When a company takes each small product decision in isolation, the temptation is always to “put it on the credit card” and pay back the tech debt later. But with better communication, planning, and tools, it is possible to reduce tech debt buildup and identify risks before they become front-page news.
Jon Stephens is the group chief product officer at Nortal, an end-to-end technology services provider.