The Trap of 'Best Practices'

“Best Practices” are great for tutorials and textbooks. They are the safety rails of our industry.

But any time I hear them mentioned at work, I feel like someone was just tired of thinking about the problem in front of them.

Early in my career, a boss gave me advice that stuck: “If you don’t know what you’re doing, copy your competition. When you learn enough to see where they’re wrong, go your own way.”

That is the definition of a Best Practice: A simplified process that works on paper, and to a limited degree in practice, but ignores all subtlety and circumstance.

Take Econ 101 and the standard Supply and Demand graph. That’s a “Best Practice” model of the world. If price goes down, more people buy, as long as there’s a willing supplier. It’s absolutely logical and you risk looking like a clown if you argue otherwise. And yet, it doesn’t survive scrutiny if you add some circumstance. It fails to explain why you would rather buy $100 shoes than an identical-looking pair for $1.

Math says the $1 pair is the optimal choice. Psychology says the $1 pair is a trap.

The math can’t account for the feeling that something is wrong. It ignores Risk Assessment. If I have a meeting in an hour and I need shoes now, I am not buying the $1 pair. The risk of them falling apart is too high. I pay $100 for certainty. If my meeting is in two weeks, I’ll try the cheap ones, what do I stand to lose?

Best practices tend to assume rational actors in a vacuum. Customers are emotional actors in a high-stakes environment.

Work Backwards from the Result you Want

When we built the PC Builder at Corsair, the “Best Practice” for web performance was clear: Speed is God. Every millisecond of latency kills conversion. Google Core Web Vitals demands instant rendering.

We deliberately broke that rule.

We added an artificial loading screen. We made the wizard pause between steps. We forced the user to wait while the system “calculated compatibility”, “tabulated power draw”, “evaluated thermal requirements”, and “made sure it fist in the box”.

Why? Because buying PC parts is terrifying for a novice, and we were developing a system that would help them deal with the complexity. If the system returns a result instantly, it feels like a database lookup, a marketing scam that pretends to do the math. It feels insincere and cheap. But if the system “thinks” for a moment, it feels like an engineer checking your work.

That artificial friction didn’t hurt conversion, it increased trust. And trust was what we were selling.

And because life can’t just be simple, here is where some more circumstance comes in.

My initial idea was to keep making the “calculation” pause longer until we see analytics tip over and users leave because of the delay. But there were some users who weren’t there for the trust and confidence.

I had a colleague who had to test and maintain that system every single day. If I forced him to sit through a 4-second fake loading screen 250 times a day, he was going to lose his mind.

We compromised around 1.7 seconds.

• The Best Practice said “Make it instant.” (Wrong for the customer).
• The Pure Theory said “Make it 4+ seconds.” (Wrong for the team).
• The Contextual Strategy landed around 1.7 seconds.

Learn the best practices so you understand the rules. But mastery isn’t about following them. It’s about improving your understanding of the problem, and knowing which rule to break to solve it.