Velocity theatre and the people who measure the wrong thing

We had done it. We deployed when we wanted, with full confidence. The team owned its work. We sliced thin and counted what shipped.

Then the organisation started measuring us again. Not flow. Not value. Velocity, and people. This is how a good team gets eroded.

Every ticket is a three

We stopped estimating.

For years I had watched teams burn hours debating whether a ticket was a three or a five, as if the number meant anything once the work started. It was waste, and it did not represent value. So we made every ticket a three. One small, uniform unit. Slice the work thin enough and the size stops mattering. You just count what you ship.

This is where the worldview from building the team became a practice. Jeff Patton taught me to slice thin vertical slices of real value. Marty Cagan taught me to chase outcomes over output. Put those together and the conclusion is simple. Make tickets work for you, not the other way around.

So we showed DORA instead of burndown.

Lead time. Deployment frequency. Change failure rate. Recovery. Those told us whether we were getting better at shipping valuable change safely. A burndown chart told us how good we were at filling a chart.

Gaming velocity

Then our scrum master left, and we got a replacement.

He was lovely. He also wanted traditional velocity. He wanted the chart, the points, the predictable line.

So we gamed it.

One sprint we did far more than usual. The next we did half. We made the numbers swing until they were meaningless. He saw what we were doing, understood the point, and stepped back to let us work the way we worked.

That sounds like a petty trick. It taught me something real. A metric you can game that easily is a metric measuring the wrong thing. Velocity measures activity. It says nothing about whether the activity was worth doing. We could move that number in either direction without changing a single outcome for a single user.

Then they hired Ruth

Nobody on the team was asked. None of us met her before she arrived.

Ruth took control of our meetings. Where people used to speak naturally, in no particular order, building on each other, she made us go round in turn. When a conversation reached the detail that actually mattered, she would say "take it offline," at exactly the wrong moment. The safe space we had built dissolved. Our free flow became regimented.

She was not what we needed, and not what we had asked for.

The clearest tell was the language. Ruth referred to people as resources.

We never did. We saw individuals with unique strengths, weaknesses, and interests, and we designed the work around them. People first, then product. We knew who loved which part of the system, who would watch every talk on a topic, who would read everything and raise the whole team. That was our edge.

Someone argued this created silos. I argued the opposite. Rotating people around as interchangeable parts is the real tell of treating people as resources. When you are lucky enough to work with people who are very good, and who love what they do, you do not flatten them into a staffing spreadsheet. You let them go deep. A team of specialists who care will beat a team of interchangeable units every time. We chose good over mediocre and even.

Honesty and transparency were core values for us. When the hierarchy wanted a private conversation about Ruth, I stopped it and asked that she be present. What is said about someone should be said openly, with them in the room. That was the standard we held ourselves to.

What it cost

Looking back, Ruth was not the disaster. She was the first sign of one.

This was the first of many steps Cambridge University Press took, from being a great place to one that lost its magic. Key people left, including me. The company turned to outsourcing to cheaper teams in Poland. The thing that had made us good, people who cared, given the room to care, stopped being the thing the organisation valued.

It was not Ruth's fault. She was the wrong fit, dropped into a team she was never introduced to, asked to impose order on something that was already working. The failure was structural. The organisation reached for control and standard process at the exact moment it should have protected what was working.

The lesson

The model worked. That part is not in doubt.

We proved you can deploy on demand, with confidence, with low rework. We proved that small teams, real ownership, and feedback pulled early beat a heavy pipeline every time. The DORA numbers moved because the system changed, not because anyone went faster.

What I could not protect was the team and the way of working.

Organisations keep measuring people and output instead of flow and value. They count velocity because it looks like productivity. They call people resources because it makes a spreadsheet simpler. And in doing so they erode the very things that made them deliver in the first place.

That is the warning I would leave you with. You can build something great. You can have the numbers to prove it. None of that protects it from an organisation that has decided to measure the wrong thing.

It worked, while it lasted.