Why Your Success Metrics Are Lying to You (And Mine Were Too)

Last month, I shipped a feature that hit every metric we'd defined as success. Load time under 2 seconds. Zero accessibility violations. Clean code. The product manager was happy. My team celebrated.

Then a user complained that the feature was designed specifically to make them spend more money. They weren't wrong. We'd optimized for engagement and conversion, which meant dark patterns were baked into the UX from the start. The metrics never caught it because we'd never defined "ethical use" as something to measure.

That conversation stuck with me, and when I read about redefining success metrics in design, it felt like someone had finally named the thing I've been frustrated about in every job I've held.

The Illusion of Balance

Most of us learned about the classic success triangle: desirable, feasible, viable. Three equal corners. Looks nice on a slide.

The honest version? Viability eats the other two for breakfast. Money decides whether your product ships. Desirability and feasibility are just the means to that financial end. In practice, this means I've watched genuinely user-centered ideas get killed because they didn't generate enough revenue velocity.

I've been complicit in this too. When a client asks for something I know is manipulative but profitable, my first instinct isn't always to push back hard enough. The budget is real. The paycheck matters. The ethical objection feels abstract.

The Real Lever Is Higher Than You Think

Here's what actually changed how I think about this: I was trying to solve the problem at the wrong level.

I'd attended workshops on ethical design. Read articles. Added accessibility checks to my development workflow. Checked boxes on dark reality sessions. All of it helped marginally, but none of it addressed the actual system steering the ship.

The insight that landed for me was this: you can't ethically design your way out of a business model that demands infinite growth and doesn't measure ethics. You're optimizing for the wrong thing.

This means the real work isn't in my design process or my code. It's in changing what success actually means to the organization. Not as a nice-to-have value statement. As a measured, tracked, rewarded objective.

What This Looks Like in Practice

I'm not going to pretend this is easy or that I've solved it. But I've started experimenting with how this actually works:

At my current shop, we started tracking three metrics alongside our usual conversion numbers: time spent on the platform (which we now actively want to be lower), user agency (can someone easily quit or modify their choices?), and a privacy score based on data collection minimalism.

These don't override business metrics. But they're in the dashboard alongside revenue. When leadership sees that we're slowly improving user agency while maintaining growth, the needle shifts. Slowly.

The second thing I've done is stop waiting for permission. I build small experiments into sprints that test whether we can achieve business goals differently. Sometimes they fail. Sometimes they work better than the original approach. Either way, we're gathering data on what's actually possible, not just what we've always done.

What I'm Still Wrestling With

Here's what I don't have figured out: how do you do this at a startup that needs to survive? I work with enough early-stage founders to know that the pressure is real and the margins are thin. Telling a founder that their growth targets are unethical before they've hit product-market fit feels naive.

But then again—do we build the unethical stuff now, promise to fix it later, and hope later comes? That almost never happens.

I think the honest answer is that this requires choosing which organizations you work for. Which sounds privileged, and maybe it is. But I've started turning down projects where the core business model demands harm. That's my version of changing the system: voting with my labor.

Your Move

The core challenge in that original article haunted me: profit is just a measurement. We've decided it's the only measurement that matters. But measurements are chosen. They can be changed.

If you're building something right now, what would your success metrics look like if you actually measured what you claimed to care about? What would have to change in your organization for ethical design to show up on the dashboard next to conversion rates?

I'm genuinely curious how you're thinking about this in your own work. Are you running experiments? Pushing back? Complicit? I've been all three in different projects.

Source: This post was inspired by "Designers, (Re)define Success First" by at A List Apart. Read the original article