Merging two design teams is a leveling problem

After TheoremOne was acquired, I co-led the integration of its design team with Zemoga's into a single 200-person global practice at Monks. I've mentioned before that it closed in under nine months, faster than any other function, and used that as evidence for something else. This is the part I skipped: what the work actually consisted of.

Almost none of it was the org chart.

Two ladders, one word

Both companies had senior designers. Neither meant the same thing by it.

That sounds like a naming problem and it is not. Every consequential decision in a merged org runs through the level: who gets staffed on which client, who reviews whose work, what someone is paid, who is considered for the lead role that just opened. Until the two ladders are reconciled, none of those can be made — or rather, they get made anyway, by whoever is in the room, using whichever company's definition they grew up with. That's how a merger quietly becomes an annexation.

So the first real deliverable isn't a structure. It's a single definition of each level, specific enough that two managers from different companies reading the same portfolio land in roughly the same place. Then everybody gets mapped onto it, which is slow, uncomfortable, and the thing everything else waits on.

Speed is a retention strategy

The reason to do it fast is not efficiency. It's that the ambiguity is what costs you people.

In my experience the ones who leave during an integration don't leave because the answer went against them. They leave during the months when there is no answer — when nobody can say what their title will be, who they'll report to, or whether the ladder they've been climbing still exists. A person told "you're mapped to this level, here's the reasoning, here's what the next one requires" can decide whether to stay. A person told "we're still working through it" is being asked to wait on someone else's timetable with their career as the stake, and the good ones don't have to.

That reframes what you're optimising for. A leveling system that is 80% right in four months does more good than one that is 95% right in a year, because the second one loses people during the eleven months it takes to be careful.

What doesn't merge on that schedule

Ladders, career paths and hiring operations are tractable — they're artefacts, and artefacts can be written down and argued about. Craft culture is not.

How each team decided things — what got escalated, how much divergence was tolerated in a critique, whether a lead's opinion settled a question or opened one — those are habits, and they don't respond to a document. They take much longer, and pretending otherwise by declaring a merged "way of working" in month three produces a document everyone nods at and nobody uses.

The honest sequencing is: reconcile the ladder fast because it unblocks decisions and stops the bleeding, then let the working culture converge slowly through people actually working together on real engagements. Trying to do both at merger speed gets you a worse version of each.

What I'm not claiming

I co-led this, not led it, and the nine-month figure is the merger closing rather than everything being finished. I ran a 100-person product design org inside the 200-person practice, so parts of this are observed rather than mine. And I have no retention numbers to offer — what I have is the pattern of who left and when, which is evidence of a weaker kind and worth labelling as such.

Related: scaling a design team, and how I work.