Aug 17, 2026
Visibility is an engineering problem too
Why I became more interested in making good work visible, and how that eventually influenced the tools and team structures I built.
One thing I underestimated as an engineering leader was how much organizational visibility affects people’s careers.
I used to think good work would eventually speak for itself.
The longer I managed teams, the less convinced I became.
The problem with memory
Leadership teams rarely see everything that happens inside an engineering organization.
They see major releases.
They hear about incidents.
They hear about projects that are running late or projects that went particularly well.
They see people presenting in meetings.
But a lot of engineering work happens outside those moments.
A developer who repeatedly helps other engineers may not have a large feature attached to their name.
Someone who keeps a difficult system stable may not have a launch to point to.
Someone who improves the quality of a codebase can spend months making things better without creating a dramatic before-and-after moment.
When decisions are made from memory, the most recent and visible events tend to have an advantage.
That’s where recency bias becomes an organizational problem.
I saw this in recognition first
This is where I first noticed the pattern clearly enough to name it. We’d built a recognition process, and the people doing the most visible work were naturally the easiest to remember when it came time to decide who’d performed well. That didn’t mean the recognition was wrong. It meant our view was incomplete — I wrote about what we actually changed here.
What stuck with me afterward was that recognition wasn’t the only place this showed up. The same blind spot can affect performance conversations, promotions, staffing decisions, even deciding who gets pulled into an important project.
Pods helped, but for a different reason
When we moved toward a pod-based structure, one of the goals was better visibility into what teams were actually doing.
The point wasn’t to create more meetings.
It was to make ownership clearer and make progress easier to understand without requiring everyone to constantly report upward.
With multiple products running at different stages, I needed a way to understand where work was moving, where it was stuck, and where people were taking ownership.
A good team structure reduced some of the distance between the work and the people responsible for understanding it.
This eventually became Ladderline
The same question stayed with me after I moved back into hands-on building:
What if useful evidence didn’t have to be reconstructed at the end?
That became one of the ideas behind Ladderline.
Instead of trying to remember six or twelve months of work when a promotion cycle arrives, capture evidence as the work happens.
Not a diary.
Not a system that decides who deserves a promotion.
Just a record of what happened, grounded in the work itself.
The idea came from the same frustration I had seen inside teams: important work is easy to lose when the organization only looks at what is most recent or most visible.
Visibility isn’t the same as surveillance
More visibility doesn’t mean more status reports.
It doesn’t mean asking engineers to explain themselves every day.
And it doesn’t mean turning every activity into a metric.
The goal is to make meaningful work easier to understand without creating another reporting burden.
That is also why I became more interested in systems that capture evidence naturally rather than asking people to reconstruct it later.
What I changed my mind about
Earlier in my career, I thought the manager’s job was mostly to know what was happening.
Later, I started thinking the better goal was to build systems where a manager didn’t need to know everything personally.
The organization should make important work visible.
Teams should have enough ownership that progress is understandable.
And when decisions need to be made, the information should come from closer to the work rather than only from whoever happens to remember it most clearly.
I don’t think recency bias can be eliminated.
But good systems can make it harder for recency to become the only thing people remember.
← Back to writing
Comments
Loading comments…