Track Quality After You Ship:One Crude Number Is Enough
How to track quality after you ship: pick one crude signal, define it once and read it beside your speed. Here is how to spot work moving into the future.
To track quality after you ship, pick one crude signal and watch it next to your speed: things that broke after release, time spent redoing work, or changes that needed an instant follow-up fix. Precision matters less than consistency. Without it, you can't tell progress from debt.
How to track quality after you ship, without building a system
You track speed. Maybe not formally, but you know when you shipped. You feel it.
Quality is different. Nothing in your week tells you it slipped. There's no invoice for it. No colleague frowns. It leaks out slowly, and you find it months later as "everything takes longer now".
That's the problem. When you work alone, nobody else is watching the second half of the equation. You are the builder, the reviewer and the support desk. If you don't count it, it doesn't exist.
So here is the verdict. A speed number on its own is worse than no number. It rewards you for shipping more, and shipping more is exactly how work gets quietly pushed into next quarter. You shipped faster. You also redid more. That is not speed. That is borrowing.
The fix is small. To track quality after you ship, keep one rough quality signal beside your speed signal, looked at on the same day. That's it. No tooling project, no dashboard.
Track quality after you ship: why speed alone lies
Speed and quality trade against each other all the time, and almost never in a way you notice. One small decision at a time. Skip the check. Accept the first answer. Ship it, fix it later.
Each decision is reasonable. The pile of them isn't.
Now add AI. It produces plausible work faster than you can evaluate it. That is wonderful for output and dangerous for judgement, because plausible and correct look identical at a glance. The trade between speed and quality now happens faster than it used to.
Here is the pattern to fear: a month that looks productive and a month that is quietly building debt feel the same from inside. Same busy evenings. Same sense of motion. The difference only shows up later, when you reopen something and it fights you.
A quality signal makes the trade visible while it is still cheap. That is its entire job. It is not there to grade you. It is there to tell you, in plain numbers, when you are moving work into the future.
Free guide · 21 pagesMeasure the system, not yourself.The Delivery Metrics Playbook: is the machine healthy, without the stick.Get the free guide →Three signals you can start on Monday
You don't need all three. Pick the one you can count without effort. If counting it takes an afternoon, you picked the wrong one.
| Signal | What you count | What it tells you | Where it breaks |
|---|---|---|---|
| Broke after shipping | Things that failed or embarrassed you once they were live | Whether your release habits are catching problems before customers do | You must decide once what counts as broken, then stick to it |
| Time spent redoing | The share of a week spent on rework rather than new work | How much of your output is real and how much is repair | Needs a rough honest guess at the end of each week |
| Immediate follow-up fixes | Changes that needed a second change within a day or two | Whether you are shipping before something is actually done | Small tweaks can inflate it, so count only fixes to mistakes |
Notice what's missing from a way to track quality after you ship. There's no score out of ten. No average satisfaction. No survey. Those feel rigorous and take ages to build, and you'd abandon them within a month.
A crude number you actually read beats a clever one you don't. Every time.
Define it once, then leave it alone
The only real discipline here is the definition. What counts as "broke"? What counts as "done"?
Write it down. One sentence is enough. "Broke means a customer saw it or I had to fix it within a week of release." Whatever. The wording matters less than the fact that it doesn't change.
Because in three months you will define it differently, and then the trend will look strange, and you'll either panic or stop trusting it. Both outcomes kill the signal. Your future self can't read your current self's mind. Leave a note.
The same goes for what "done" means. Done when it works on your machine? When it's live? When a real customer can actually use it? Pick the one that matches what reaches a person who pays you. Anything earlier flatters you.
Read it beside speed, not instead of it
Here is where the signal earns its keep. You look at both on the same day, and you ask one question: did one move without the other?
- Speed up, rework flat. Good. You got faster for real.
- Speed up, rework up. You didn't get faster. You moved the work later.
- Speed flat, rework down. Quietly excellent. Nobody will congratulate you. Congratulate yourself.
- Both worse. Something structural is off. That's a flow or capacity question, not a quality one.
The second line is the one that fools people. It looks like success for about two months. Then a particular area of the work starts to slow down everything near it, and you can't say why.
If you want to go one step deeper on diagnosing that slowdown, how to tell a capacity problem from a flow problem is the next read. Quality signals often point straight at it.
Watch for the cliff, not the slope
Quality decay doesn't arrive as a gentle decline. You get a stable stretch, then a sharp jump in how long everything takes and how often a change breaks something else. By the time it's obvious, it's expensive.
The warning signs move earlier, though. Three to watch:
- A rising share of your time going on rework.
- One area of the work getting slower than everything around it.
- More changes that need an immediate follow-up.
Those move months before the cliff. That's why a crude signal is enough. You don't need to measure the drop precisely. You need to notice the bend.
And when one area keeps showing up, you now have something better than a feeling. "Anything I touch there takes several times longer and breaks more often" is an argument. It lets you choose, on purpose, to fix the unglamorous thing instead of building the next shiny one. Without the number, the fix always loses. The new feature is visible. The repair isn't.
Free assessmentHow trapped are you, really?Take the Corporate Suffocation Index, the free solopreneur readiness assessment.Take the assessment →Check the number against your month
Before you trust any of it, hold it up against how the month actually felt.
If you felt like you were firefighting and the signal says all is calm, your counting is wrong. Maybe problems got fixed quietly before you logged them. Maybe work is happening outside whatever you track. The data is giving you a bug report about your measuring, not a revelation about your business.
Treat it that way. Fix the counting. Don't start doubting your own experience, because that's how people end up distrusting the whole exercise and dropping it.
What not to do with this number
Don't set a target. "No more than this many things broken" sounds sensible and it is an instruction to hide breakage. There is always a way to make a number look good without improving anything. You will find it, probably without meaning to.
Don't compare yours to anyone else's. Their definition isn't yours.
And don't use it as a verdict on yourself. A bad month usually says something about the system, or about a month that had other things in it. Read it as information. If you can't, don't track it. Building unmeasured is better than building while quietly demoralised by your own spreadsheet. There's more on that line in delivery metrics mistakes to avoid.
Free guide · 21 pagesMeasure the system, not yourself.The Delivery Metrics Playbook: is the machine healthy, without the stick.Get the free guide →Your first week, concretely
- Pick one signal from the table. The one that costs you the least effort to count.
- Write its definition in one sentence and put it somewhere you will find it.
- Set a recurring slot, once a week or once a month, when you look at it next to your speed.
- Ignore it between slots. Especially on a bad night.
After a month of this you'll have a baseline for how you track quality after you ship. After two you'll know what an ordinary bad week looks like, so you can spot a real one. That's when it starts paying.
Free assessmentHow trapped are you, really?Take the Corporate Suffocation Index, the free solopreneur readiness assessment.Take the assessment →Frequently asked questions
What counts as a quality problem for a one-person business?
Anything that broke or needed redoing after you called it done. Pick a definition, write it in one sentence, and keep it. Consistency matters far more than getting the definition perfect.
Do I need special tools to track quality after you ship?
No. A note you update at the end of each week is enough. A crude number watched steadily will tell you when quality is being traded for speed, and that's the whole purpose.
Why does AI make this more important?
AI produces plausible work faster than you can review it. Without a quality signal, a productive month and a debt-building month are indistinguishable until later, when the cost has already grown.
Should I set a quality target?
No. A target tells you to make the number look good, and there is always a way to do that without improving anything. Watch the trend and investigate movement instead.
The full method, with the checklists, is in The Delivery Metrics Playbook, a free 21-page playbook.
How trapped are you, really?
Take the Corporate Suffocation Index, the free solopreneur readiness assessment.