One operating model, many sites, none of them the same.

Standardizing the way many locations work is the purest test of whether a change can actually hold. It is also where good intentions most reliably go to die.

Leadership wants every site running the same playbook. It makes sense on paper. It almost never survives contact with the sites, and understanding why is the whole game.

I have seen more standardization efforts fade than nearly any other kind of change, and they fade for reasons that are entirely predictable once you have watched it happen a few times.

Why it keeps failing

Every site has its own habits, its own informal leaders, and its own genuinely good reasons that the local way works. A standard handed down from the center, without local buy-in, reads as an outsider telling people who run the place day to day that they have been doing it wrong.

So it gets a polite nod and quiet non-compliance. The mandate lasts about a quarter. Then each site drifts back to its own version, the rollout team declares victory and leaves, and a year later you have one official operating model and a dozen real ones.

Nobody refused. Everybody just quietly kept doing what worked for them, which is the most durable form of resistance there is.

The center underestimates the local truth

Here is the part that trips up smart leadership teams. They assume variation between sites is a problem to eliminate. Often it is, but some of it is local adaptation to real differences, and the people at the site know which is which far better than the center does.

If you steamroll all of it, you break things that were working and you lose the trust of the people you need most.

The art is telling the difference between dysfunction worth standardizing away and adaptation worth keeping, and you can only do that by involving the sites rather than dictating to them.

Standardize the what, allow some how

The standardization efforts that last tend to draw a clear line between outcomes and methods. The outcomes, the what, are non-negotiable and shared across every site. Some of the how can flex to local reality, within bounds.

A standard that dictates every keystroke will be resented and ignored. A standard that fixes the outcomes and the core process, while leaving room for sensible local variation, gets adopted because it does not insult the people running the work.

Knowing which parts to lock and which to leave flexible is the central design decision, and getting it wrong in either direction kills the effort.

Architect once, adopt everywhere

The pattern that works is to design one repeatable operating model, then roll it out in a sequence that earns belief rather than demanding it.

Prove it at a site or two first. Pick locations where it can win, make the result visible, and let that result do the persuading the mandate never could.

When other sites see peers succeeding with the new way, the conversation changes from why are you forcing this on us to how do we get what they have. That shift is worth more than any directive from headquarters.

Peer proof beats central mandate

A site will discount anything that comes from the center and trust almost anything that comes from a peer site like theirs. This is not stubbornness. It is a reasonable judgment that the center does not fully understand their reality and a peer does.

So the most powerful tool in a multi-site rollout is not a better mandate. It is a lighthouse site, a comparable location that adopted the new model and is visibly better off for it, whose people will tell the truth to their peers.

Spend your energy creating that proof, and let it travel through the network on its own. It moves faster and lands harder than anything you could send down the org chart.

Local owners, not just central mandates

A standard that lives only at headquarters does not survive in the field. Each site needs a local owner who is bought in, credible to their peers, and accountable for keeping the standard honest after the rollout team is gone.

Central defines the model. Local makes it real. Skip the local ownership and you are relying on a mandate to hold behavior in a place you do not work, which is exactly the bet that keeps failing.

The local owner is not a compliance officer. They are a believer with standing, which is a very different and far more durable thing.

Anchor it or it drifts

Standardization is never done at launch. It needs a cadence that outlasts the rollout team: shared metrics that every site sees, regular review, and reinforcement that keeps the model from quietly fragmenting back into local variants.

Without that anchoring, entropy wins. Every site drifts toward its own version, slowly enough that nobody notices until the standard is a memory.

The work is not getting the sites to adopt. It is getting them to stay adopted, and that is a permanent discipline, not a one-time event.

If you are rolling one model across many sites and it keeps fading, the problem is rarely the model. It is that you are mandating where you should be proving, and dictating where you should be transferring ownership.

Win a site, make it visible, give each location a real owner, and anchor it with shared measurement. That is how a standard actually sticks across a network, instead of dissolving back into the dozen versions you started with.