UK investment platform · Financial services
Engineering management and agile delivery across 12 squads
A UK investment platform had outgrown its engineering structure, with too few managers for too many engineers. I built its engineering management practice, then took 12 squads through an agile transformation. Retention and engagement rose, and engineers were promoted into the new manager roles.
- 12 squadsand 60 to 100 engineers, moved to flow metrics and forecasting
- From withinengineers promoted into the new engineering manager roles
- Retention upand engagement scores rose with it
Where they were
The client is an FCA-regulated UK investment platform with hundreds of thousands of customers. Its engineering organisation was growing, and the shape that had worked for a smaller team no longer fitted. Engineers reported to a handful of senior leaders, each with far more direct reports than anyone can look after properly.
In my experience that arrangement fails slowly. One-to-ones get cancelled, career conversations get put off, and good engineers start wondering whether anyone is invested in them.
I worked with them through More Than Monkeys for more than a year: first designing and running their engineering management practice, then leading an agile transformation across twelve squads of between 60 and 100 engineers.
Managers first
The order mattered. Changing how twelve squads deliver needs someone in each team who owns how that team works and how its people grow. Without them, every new practice ends up depending on the same few overloaded leaders.
So I started by defining what an engineering manager does there: the responsibilities, the levels, and how someone progresses from one to the next. Then came the people. The roles were filled partly by promoting engineers from inside the squads, and I coached the new managers, with a regular rhythm of one-to-ones, reviews and coaching around them.
I’ve written more about the role itself in how to be a software engineering manager without turning into a middle manager.
Squads built around what they own
With managers in place, the transformation started with the shape of the teams. Each of the twelve squads got a clear area of ownership, a reporting line that made sense, and a product partner to work alongside, so the people deciding what to build and the people building it were in the same team.
Forecasting from real delivery data
The squads moved from story-point estimates to flow metrics: cycle time and throughput, measured from work they had actually finished. Delivery forecasts came from that history, as a range with a likelihood attached, which is a far more honest answer to “when will it be done?” than a velocity chart.
The reasoning behind that is in Story Points Have Become a Joke, and the method is in Probabilistic Forecasting.
What changed
Retention improved, and engagement scores rose with it. Engineers could see what progression looked like, and said so. Some of the new engineering managers came from inside those same squads.
The squads had a way of working they owned, and the organisation had a management layer built to carry it on without me.
Why it worked
In my experience, most agile transformations start with ceremonies and tooling and leave the management structure alone. The new process then sits on top of the same overloaded people, and fades soon after the consultants leave.
Fixing the management layer first meant every squad had someone whose job was to make the new way of working stick, and to grow the people doing it.