Capital loves code. Over the past decade, there’s been no better way to get fat returns on your cash than to invest in code. Tech stocks have buoyed markets. And code fever shows no signs of abating.
But the cool new code we write today becomes the massive technology debt of tomorrow, and there’s nothing sexy or profitable about debt service.
Don’t believe me? Just visit the data center of any global financial institution. There you’ll find a mainframe running COBOL programs from the 70’s. That’s right, the 70’s. You know, that time before computers.
Now multiply the technology debt from the 70’s by a few orders of magnitude. Think of all the code we’ve written in the intervening decades and all the places it’s running. Add to that all the code we’re writing now and will write in the coming years. The apps, the databases, the AI.
It’s unsustainable. It’s not an immediate pain-point. So like climate change, infrastructure, and all our other long-term challenges, technology debt is a can that capital just keeps kicking down the road.
Meanwhile, we keep producing massive future technology debt at an unfathomable rate. Worse yet, we encourage young people to become coders and add to the burden.
A lot of this code is bad. A lot of it is unsecure. And all of it – 100% -- is future debt.
Technology debt is not an abstraction. It has a real cost. You have to allocate time, money, and talent to old code. Or, alternatively, you can ignore it – in which case you pay for your neglect with poor business performance, non-competitive customer experience, and outages.
Technology debt is why you pay Microsoft a tax just for owning a CPU and having to run Word. It’s why the IRS shuts down just to modify a few pieces of business logic. It’s why Walmart had to spend $3.3 million on Jet.com.
Of course, no one thinks they’re creating debt when they roll new code into production. After all, their code is brand spanking new. It has been birthed by state-of-the-art code cultures with all the right attributes: Agile, DevOps, Continuous Delivery.
But it’s still all debt. And you’ll never re-platform your way out of it. Heck, I’m writing this on a platform that dates from the Reagan administration.
Can we avoid technology debt entirely? Of course not. But we can:
- Price debt into the code we produce today – instead of persisting in our denial-based valuation of intellectual property.
- Better service the technology debt we already have, including our mainframe code – which is some of the most important code on the planet running on some of the most awesome machines on the planet.
- Stop being so juvenile and manic about coding. Not every kid needs to be a coder. We need plumbers and welders, too.
In fact, maybe we should ease up on coding for minute while we address what is perhaps our most egregious “technology debt” – the roads and bridges and airports and water/sewer systems we have neglected for decades.
We can’t code our way out of debt. We can, however, choose to more prudently manage it.