Role: Senior Engineering Manager · Company: CarTrawler
The problem
CarTrawler’s engineering org had grown organically into a front-end discipline and a back-end discipline. Almost every feature needed both, so almost every feature needed a handoff. Work sat half-finished waiting on the other side, nobody owned a feature end to end, and cycle time was measured in weeks.
The hard part was not the org chart
The structure was the problem, not the people in it. Engineers and managers on both sides had been performing well within a setup that made delivery slow. Saying that clearly, without it sounding like blame, was most of the work.
The other risk was real: put specialists into mixed squads and you can dilute the depth that made them good, and slow everything down while people adjust.
What I did
I ran discovery sessions with engineers, tech leads and product managers to get the friction points from the people living with them. The redesign came out of that input rather than being handed down.
Then I piloted it. One product area moved to full-stack squads first. The results from that pilot did the convincing for the org-wide rollout, which I would not have got on a slide deck alone.
To protect specialist depth I set up communities of practice that cut across the squads. Front-end and back-end engineers kept a home for their craft even though they no longer sat in a discipline team.
What changed
Feature cycle time went from weeks to days. Squads reported higher ownership and a clearer sense of purpose. Attrition went down. The communities of practice outlived the reorg and became a standing part of how CarTrawler’s engineering culture works.
What I would keep
Pilot first, every time. And build the thing that preserves what the old structure did well before you take the old structure away.