Escaped Defect Rate: Quality is about hard truths, not ego boosts
A lifetime ago (we're talking school age), my friends and I would joke about me being a robot. I was always agnostic on where we went, what we did. Never fully committed to what I liked and wanted. Hedging all my emotions. Forever trying to fit myself into the system in front of me.
The thing is, the robot wasn't broken. She was functional. She integrated. She got good grades, made people comfortable, never asked for too much. Always smiled. She was easy to have around.
That was the whole problem.
QA can get bogged down on easy numbers. Thousands of test cases run. Requirements coverage achieved. Defects found in testing. None of it is a depiction of quality, usability, or value — but it's easy to point to and say, see, all that work must mean something. All those people saying you're a hard worker must mean something.
But what about the defects that get through?
Escaped Defect Rate measures the issues found in production. After testing, after sign-off, after everyone said it was ready. It puts a spotlight on all that coverage because if you ran 1000 test cases with 100% coverage and still shipped critical failures, you weren't testing the wrong number of things. You were testing the wrong things entirely.
I've been running an experiment. Tracking how often I find myself asking what I want after I've already made a choice. Noticing how often the decision came from external factors instead of internal ones.
It happens across the whole range. Small stuff like everyone ordered iced coffee so I did too, even though I wanted something warm. To large stuff like my whole family lives in California, I'm over the cost of living, and I'm still here. The test cases all passed. Socially acceptable input, expected output, no errors thrown. But the defects escaped anyway.
The most expensive one I've found so far: my college degrees—plural.
I went to college in the first place because I wasn't ready to be an adult yet. I had to pick something to major in so I picked linguistics. Languages genuinely fascinated me. I love that some languages have a single word for what takes a whole phrase in English to express. I loved phonology for the pattern recognition. I had real curiosity and actual interest in the subject. Something that was mine.
But then I picked computational linguistics for my concentration because it sounded marketable. I figured the only way to make money from the BA was for it to be technical adjacent.
I took something I was genuinely curious about and immediately optimized it for someone else's requirements. Then I kept going and added a master's degree. Told people it was for career advancement and expertise. But the real reasons, if I'm being a quality advocate about it, were that I wanted to stay out of the workforce a little longer and I wanted to have more education than other people. The proof of capability mattered more than the capability itself.
The hard truth is I've never done anything in the field of linguistics—which was what I actually enjoyed. I put up with the computer science classes to justify what I wanted to learn about—languages and different cultures and points of view.
The job I got out of grad school was in customer service. I can make up narratives on how things I learned help me. That they interviewed someone without experience because of the degree. But I had a great cover letter, I interview well, I'd have gotten a job, I'd have worked my way up somewhere.
It was a milestone without a requirement behind it. Full coverage, degree conferred, all green. And in production, in the actual use of my life, it maps to nothing. The escaped defect rate on that decision is high. The failure just took years to surface.
The harder part is that it wasn't a one-time thing.
Work feels like this now too. My whole career if I'm being honest. Milestones without requirements. Get a real job. Get the promotion. Keep moving up. Give a decade at one company. Be laid off. Get the Director title. Get a good salary and design for stability. Don't question purpose or look for passion.
I've been shipping to someone else's specifications for so long I stopped noticing there was another option.
I know what the fix looks like, roughly. Work that belongs to me instead of work I trade time for in exchange for a salary and a title that look good from the outside.
But the constraints are real. I have two kids under five. I have a mortgage. I have golden handcuffs that are genuinely golden — good enough that walking away requires more than clarity of purpose, it requires a plan I don't have. The defect is logged. The priority is high. The fix is blocked.
And underneath the constraints is the harder admission. Even if they lifted tomorrow, I'm not entirely sure what I'd be building toward. I've spent so long integrating into other people's systems—their expectations, their definitions of success, their idea of what a hard worker looks like—that my own requirements are still being written. I'm only now learning to ask what passing actually looks like for me, and too often after the fact.
That's what the experiment is for. Tracking the small choices first. Ordering the warm coffee. Noticing the gap between what I reach for and what I actually want. Building an internal test suite, slowly, from the data up.
The robot integrated seamlessly into every system she encountered.
She just never thought to ask whose system it was.
Bug reports Tuesdays | Essays Saturdays
Never miss an issue.