What I think about bringing two engineering teams together

A wild horse grazing in tall grass near the dunes on Cumberland Island

When two engineering teams become one, through an acquisition or a reorganization, the conversation tends to go straight to tooling. Whose pipeline, whose framework, whose branching model. I think that is the easy part, and the part most likely to go wrong if decided first.

Protect what made the smaller team work

Two teams are rarely the same size, and the smaller one knows it is being absorbed. Everything it did well, it did with its own tools and habits, and declaring a winner in week one is the fastest way to lose those. The habits go, the people who valued them follow a few months later, and the organization keeps the tooling and loses the reason it wanted the team.

So my first move is to find out what the smaller team is proud of and protect it, visibly, before any standardization talk starts. The team should hear that its way of working is being studied, not replaced.

Share the few things that must be shared

A few standards have to be common, and I think the list is shorter than most organizations assume. How code reaches production, how changes are reviewed, how security is handled, and how incidents are run. Nearly everything else, from editor settings to the shape of a sprint, can stay different for a long time at no cost.

I would agree the short list early and let either team's version win each item on merit rather than size. A merged team that sees the larger side lose a few decisions learns that the process is real.

Put both teams on one real thing

Documents and all-hands do not integrate teams. Work does. The most effective move I know is a real piece of work with a real deadline, staffed from both sides, where people must depend on each other to ship. The friction that surfaces is the friction that matters.

Do not decide the hard things in the first month

The first month is when everyone knows the least and feels the most pressure to show progress. Decisions that are easy to make and hard to reverse, like which codebase survives, should be deliberately deferred. Say when they will be made and what will be learned first. Deferral with a date is a decision. Deferral without one is drift.

Listen for the phrase

There is one phrase I listen for in a merged team, "the way we used to." Said with affection, it is history. Said with resentment six months in, it means the integration stalled and the smaller team is still separate, sharing only an org chart. At that point the tooling decision stops mattering, because the people have already decided.

None of this is fast. Bringing two teams together is a year of work, mostly about people and only occasionally about systems. I think the organizations that get it right plan for the year and not for the migration.

Photo source: https://photos.robertstowe.com/cumberland-island