I wrote about spec drift a few months ago. About how software systems drift from their original intent — not through big pivots, but through a thousand small adjustments that each seem reasonable in the moment.
I didn’t expect to find the same pattern in my own work.
But here I am in June, looking back at something I started in January with a clear picture in my head. The picture I have now is different. Not broken. Not failed. Just… not quite what I intended.
The thing is, I didn’t notice it happening. That’s the part that bothers me.
It’s not the big decisions
When you consciously change direction, you know you’re doing it. You can track it, reason about it, evaluate later whether it was right.
That’s not what I’m describing.
I’m talking about the adjustments that don’t feel like adjustments. The week you skip a step because you’re running behind. The shortcut that becomes the new default. The “this is fine for now” that stops being temporary.
None of those moments feel significant. Combined, they redefine what you’re actually building.
The calendar always wins the negotiation
I had a principle at the start of this year: I’d only ship work I was genuinely satisfied with. Not perfect — I know better than to chase perfect — but real. Thought-through.
By March, I was measuring “good enough” against time pressure, not against that original standard.
The calendar doesn’t argue with you. It doesn’t push back. It just makes the tradeoff visible every single week. And enough weeks of the same tradeoff quietly rewrites your standard.
I don’t think this is laziness. I think it’s adaptation. Practical adjustment to a real constraint.
But adaptation at the cost of intent is still drift. Even when it feels rational.
What I actually meant to build
The clarifying question I’ve started asking is this:
If I showed the current version to the person who started this, would they recognize it as what they meant?
Not “does it work?” Not “are people using it?” Those are delivery questions.
This is an intent question.
When I applied this recently to one part of how I operate, the answer was: mostly yes. But there were two places where I’d quietly made a compromise and told myself it didn’t matter. It did matter. I just stopped checking.
One of them I could fix in an afternoon. I’d been carrying it for three months.
The drift you don’t log
Here’s the part I find genuinely strange: the drifts that cost the most are the ones that felt most justified when they happened.
“We don’t have time for that right now.” Reasonable.
“Let’s simplify this and revisit later.” Reasonable.
“This is good enough to move forward.” Reasonable.
Every single one. Reasonable.
The problem isn’t any individual judgment. It’s the accumulation. And no one logs accumulation. You just notice it one day when you look at what you’ve built and feel faintly disappointed without being able to explain why.
The small recalibration
I’m not advocating for constant retrospectives or weekly intent audits. That’s its own trap — shifting from building to managing the anxiety of building.
What I’ve added is simpler. Once a month, I pick one thing and ask: if I started this today, knowing what I know, would I design it the same way?
Usually the answer is 90% yes. But that 10% is where the drift lives.
Catching it at 10% is a small correction. A conversation with yourself, maybe an afternoon of work.
At 40%, it’s a restructure.
At 70%, it’s a rebuild — with a story you tell yourself about why it had to happen that way.
The calendar will keep winning the weekly negotiations. That’s fine. Time pressure is real. Tradeoffs are necessary.
But if you never step back to check whether intent is still alive underneath all those reasonable compromises, you won’t notice it’s gone until you’re well past the easy recovery point.
And then you’re not building what you meant to build.
You’re building what survived.
That’s a different thing.