InfoQ has a write-up of Thiago Ghisi's QCon London talk on growing a software leadership team. The practices are concrete: weekly or biweekly syncs where leads hold each other accountable for the most important things, and one-page documents that explain the why before any reorganization, so the reasoning travels with the change.
The one I think most teams skip is calibration. Managers draft the expectations for their people at the start of a cycle, and then get feedback from their peers on those drafts, so that good means the same thing from one team to the next. The payoff Ghisi describes is leads who become multipliers: giving each other feedback and making decisions together, rather than routing everything through one person.
I think calibration gets skipped because it is uncomfortable and slow. The moment it becomes visible is a promotion or a performance conversation where two managers turn out to have been measuring different things.
My view, from the challenge of leading leads rather than engineers, is that the leadership team has to agree on what good looks like before it can agree on anything else, and that agreement does not happen by itself. It is a practice, and a scheduled one.
How does your leadership team check that it means the same thing by good?
Photo source: https://photos.robertstowe.com/iguazu-falls