The Last Mile: Why Transformation Plans Die at the Office Door
Why transformation plans written at the centre fail in the offices that have to run them, and why buyers pay only for what has landed.
I have co-written global transformation plans and been part of many more, and I have been the regional Head those plans landed on. Over two decades in international services businesses, I sat at both ends of the programme. The view from the receiving end is the one most transformation leaders never get.
From the centre, the plan is elegant. The logic holds, the numbers compound, the phasing is sensible. From the office floor, the plan is one more thing arriving on top of clients, targets and month-end. A transformation plan is a working hypothesis. Every office in the footprint is an experiment that can falsify it.
A buyer pays for what landed
Last week I wrote about the numbers a buyer prices. Transformation claims get the same treatment, only harder.
In eighteen months advising private equity buyers in this sector, I have not seen a single pack that did not present transformation as underway. Savings identified. Capacity about to be unlocked. Cross-sell machinery about to switch on. Buyers have learned to sort these claims into two piles: evidenced and asserted. None of this means the plans are dishonest. Most are written in good faith. But good faith is not evidence. Claimed capacity ceilings rarely survive contact with the actual operation. Cross-sell lands at a fraction of the ambition, and over years when the deck promised quarters. Savings that depend on offices changing how they work are discounted until someone can point to an office that has actually changed.
A buyer does not pay for the plan. A buyer pays for what the offices actually do differently.
The receiving end has its own rationality
Local offices are rarely the villains of a failed rollout. The receiving end has its own rationality, and most plans are written as if it were less important.
Offices run P&Ls. They adopt what helps their number and route around what does not. I did this myself. As a regional leader I had global initiatives arriving on my patch constantly. Some I pulled in and championed. Some I pushed back on. The org chart was never the reason. The question was always whether the thing made my clients better served and my numbers better, or just made my month harder.
Then there is the sentence every programme leader learns to dread: "No, but here it is different." Sometimes it is true. I once spent the better part of a year getting twenty countries onto one standard scope of work, and in every single country the first answer was that it could not work there. The honest answer was yes and no. The local nuances were real. They were also the last mile of the road rather than the whole of it. Learning to tell a genuine constraint from a veto dressed as a fact is half the job.
And underneath it all sit the incentives. If the office is still measured and paid on the old model, the office still runs the old model, and month-end does not pause for the rollout.
What actually lands
The programmes I saw land, and the ones I landed myself, tended to share a few habits.
They split the work honestly between what must be global and what is genuinely the local last mile. Platforms, standards and shared delivery belong to the centre. I helped build a delivery centre on exactly that split, and the split is what made it work. The last mile stayed local, and the offices could see which was which.
The work that moves to a shared centre is not the last mile. The last mile is what the office does once that work has gone: the client, the onboarding, the exceptions, the local rulebook. That is also where the knowledge has to stay. If the office no longer does the work, it stops producing the people who know how the work is supposed to be done. AI will hit the centre first. It will not design the leftover office for you.
They changed the incentives before the process. Shared metrics do more to pull a footprint in one direction than any template I have seen.
They landed through people the offices rated, because granted authority tends to stop at the office door.
They sequenced for visible wins. The first office that gets measurably faster funds the credibility for the next ten.
And they kept one version of the truth, so that the offices could see themselves in the plan rather than feeling it had been done to them.
When plans are free
AI has changed the stakes here.
A credible transformation strategy can now be drafted in an afternoon. The analysis, the benchmarks, the phasing, the business case: all of it is faster and cheaper to produce than at any point in my career. Which means the plan itself is no longer the scarce asset. What is scarce now is the landing: the incentives, the trust, the local judgement, the person the office actually listens to. No model produces that part, and it is now the differentiator.
When plans are free, the last mile is the moat.
Stopping is also a decision
In one of my regional roles I inherited a large transformation programme: a technology-driven delivery model, years in the making, with a great deal already invested. It had been a reasonable bet when it was conceived. The more time I spent with it, the clearer the problem became. It had been designed around large, single-country clients. Our client base was multinational, buying across borders and expecting one joined-up service. The programme was never going to land, not because the technology was weak, but because it answered a question our clients were not asking.
I recommended we stop it, and we did. It would have been far easier to let it run. Stopping a programme that people have worked on for years is not a conversation anyone enjoys, and the people involved had done nothing wrong. The money already spent did not come back. The money we stopped spending did. It was not a comfortable decision, but it was the right one, and it taught me the test I have applied to every programme since: whether this business, with these clients and these offices, will actually do it, regardless of how good the plan looks.
That is also the buyer's test, asked without sentiment. Show me what the offices do differently today. And sometimes the most valuable decision an operator makes is admitting a thing will never land, and stopping.
Originally published on LinkedIn, 1 September 2026.
