There’s a certain allure to new technology. The startup that just raised $50 million. The platform that promises to revolutionize your workflow. The tool that everyone’s talking about at conferences. It’s tempting to believe that the right technology choice will solve problems that have resisted solution for years.
Sometimes it does. More often, it doesn’t.
The organizations that build durable operational infrastructure tend to make technology choices that would never make a headline. They pick tools that have been around long enough to have known failure modes. They choose platforms with boring roadmaps and predictable release cycles. They optimize for reliability and maintainability over feature lists and innovation narratives.
This isn’t resistance to change. It’s pattern recognition.
The Hidden Costs of Cutting Edge
New technology carries costs that don’t appear in the pricing proposal.
Learning curves are steeper than they appear. Your team has to learn new interfaces, new mental models, new ways of troubleshooting. The vendor’s documentation is incomplete because the product is still evolving. The community forums are thin because not enough people have encountered your specific problem yet. Every hour spent figuring out how to do something basic is an hour not spent on actual work.
Integration is harder than promised. New platforms often don’t play well with existing systems. APIs are incomplete or unstable. Data formats don’t map cleanly. The "seamless integration" described in the sales demo turns into a six-month project with custom middleware. A study by MuleSoft found that integration challenges consume an average of one-third of IT budgets, and that percentage rises with newer, less-established platforms.1
The product changes under your feet. Startups pivot. Features get deprecated. Pricing models shift. The tool you selected based on a specific capability might lose that capability in the next major release because the vendor decided to focus elsewhere. You’ve built workflows around assumptions that no longer hold.
Support is a question mark. When something breaks at 2 AM, who answers the phone? Established vendors have support infrastructure, escalation paths, and SLAs with teeth. Newer vendors have a Slack channel and good intentions.
What "Boring" Actually Means
Boring technology isn’t bad technology. It’s technology where the surprises have already been discovered.
When you choose a database that’s been in production at thousands of companies for fifteen years, you’re choosing a tool whose edge cases are documented, whose failure modes are understood, and whose performance characteristics are predictable. When something goes wrong (and something always goes wrong), you can find the answer on Stack Overflow in five minutes because someone else encountered it in 2019.
Boring technology has ecosystems. It has consultants who specialize in it. It has integration patterns that are proven. It has migration paths when you eventually need to move to something else.
Dan McKinley, an engineer who worked at Etsy and Stripe, coined the phrase "choose boring technology" in an influential essay. His core argument: every organization has a limited capacity for innovation and complexity. Spending that capacity on technology choices means you have less capacity for product innovation, operational improvement, or market expansion.2 The database isn’t where you want to be taking risks. The database is where you want certainty, so you can take risks elsewhere.
When New Technology Makes Sense
This isn’t an argument against ever adopting new tools. Sometimes new technology genuinely solves a problem that couldn’t be solved before. Sometimes the old tool is so outdated that the maintenance burden exceeds the switching cost. Sometimes a new platform offers capabilities that create genuine competitive advantage.
The question is whether you’re choosing new technology because it solves a specific, validated problem, or because it’s exciting.
A useful filter: Can you articulate, in concrete terms, what this new technology will enable that your current tools cannot? Not "it’s more modern" or "it’s what everyone’s moving to," but a specific capability gap that this tool fills. If you can’t answer that clearly, you might be buying novelty rather than solving a problem.
Another filter: What’s the cost of being wrong? If a new marketing automation tool doesn’t work out, you can switch. If a new ERP system doesn’t work out, you’re in for years of pain. The higher the switching cost, the more you should bias toward proven options.
The Unsexy Path to Operational Excellence
The organizations with the most reliable operations rarely have the most sophisticated technology stacks. They have stacks that are understood, maintained, and appropriate for their actual needs.
They resist the temptation to rebuild things that work. They invest in documentation and training rather than constant retooling. They evaluate new technology with skepticism rather than enthusiasm. They ask "what could go wrong?" before they ask "what could this enable?"
This approach doesn’t generate conference talks or vendor case studies. It generates organizations that work: consistently, reliably, and without drama.
The case for boring technology is ultimately a case for humility: acknowledging that your competitive advantage probably doesn’t come from your technology choices, and that the best technology is often the technology you don’t have to think about.
Citations
1 MuleSoft, "Connectivity Benchmark Report," 2023.
2 Dan McKinley, "Choose Boring Technology," 2015. Available at boringtechnology.club.