The six-month problem
Most of what gets fixed quietly unwinds. Not dramatically, and not through any single failure. It just drifts, one small compromise at a time, until the operation looks a lot like it did before anyone intervened.
A series of operational pain points are addressed. A consultant, or an internal project team, redesigns a process, cleans up the data and rolls out new training. The numbers start to move and everyone is pleased. The engagement ends and the team moves on to the next priority.
Six months later, someone asks what happened to that improvement, and the honest answer is that most of it quietly unwound.
BCG's own conclusion is explicit: the technology is rarely what determines success. The people dimension, the organisation, operating model, processes and culture, usually is. We even see the same effect play out on television. Gordon Ramsay's Kitchen Nightmares has featured somewhere in the order of 90 to 96 struggling restaurants since 2007, and by recent counts, more than 60 of them have since closed, a failure rate north of 65%, despite every one of them receiving free expert diagnosis, a fully funded renovation and hands-on retraining. Many had simply undone the changes after Ramsay left. The fix worked and the business had the resources to sustain it. What it lacked was the will to keep operating the new way once the person who designed it walked out the door.
That is an inertia problem, and it is the single biggest reason operational improvements fail to hold. Every operation has a natural resting state, the way things get done when nobody is actively pushing against it. That resting state was never designed, it represents the accumulation of habits that have developed shift by shift, workaround by workaround, until it became simply "how we do things here." It carries enormous mass, not because it is optimal, but because it requires no ongoing effort to maintain. An improvement is a deviation from that resting state, and like any object deviating from a resting state, it needs a continuous input of energy to stay there until it becomes the new norm. The moment that energy is withdrawn, the operation begins drifting back unless something inside the business has enough force of its own to hold it in place.
This tendency has a name outside of operations too. In Clear Thinking, Shane Parrish describes what he calls the inertia default, one of four automatic mental patterns that quietly shape decisions before any deliberate reasoning gets involved. The inertia default pushes people to maintain the status quo, and Parrish notes that starting something new is hard, but so is stopping something already underway, meaning people resist change even when the change is clearly for the better. This is not to be mistaken for a failure of intelligence or character but instead is a default setting. The operations that hold their improvements are the ones that have built something strong enough to override it.
Four forces, and a fifth complication
Inertia does not operate as a single force. It shows up across several distinct dimensions of an operation, and a handover that only accounts for some of them will still fail against the ones it missed.
Environmental inertia
The pull exerted by the physical and systemic conditions surrounding a process. This includes the layout of the space, the configuration of the systems and the habits of adjacent teams who were never part of the change. A redesigned pick path will drift back toward the old one if the shelving and system defaults still quietly reward the old route. This form of inertia wins by default. It does not need anyone to actively oppose the change, it only needs the environment to keep making the old way slightly easier.
Enterprise inertia
The pull exerted by the business's structures, incentives and priorities. If a team is still measured or rewarded according to the logic of the old process, they will rationally drift back toward serving that logic, regardless of what the new documentation says. A warehouse that redesigns its exception handling but leaves supervisor bonuses tied to raw throughput will watch supervisors quietly deprioritise the new steps the moment volume gets tight.
Cultural inertia
The pull exerted by what a team believes, collectively and often unconsciously, about how things really get done. It is the quiet scepticism that says "this is how the consultant wanted it, but here's how we actually do it." It is the senior team member whose informal authority outweighs the documented process and who reverts to prior practices first, taking others with them.
Personal inertia
Quieter still, and it acts on individuals rather than the system. Nearly every operation has someone who has built real professional identity around being the one who solves problems when things go wrong. This is the person who knows the workaround and keeps the operation running through the gaps in the formal process. This person is rarely a marginal performer. They are often one of the higher performing, higher potential operators in the business, and an improvement that removes the need for their intervention, however positive for the operation, can quietly erode the thing that made their role matter. Left unaddressed, this does not usually produce open resistance, it produces disengagement, and often the eventual departure of someone the business could not afford to lose.
The same profile that creates this risk is also the operation's best asset for preventing it. Someone with the judgement, credibility and hands-on knowledge to solve problems nobody else can is precisely the person best positioned to become the internal champion for the new way of working, if they are given that role deliberately rather than left to feel replaced by it. The difference between losing this person and gaining a genuine advocate usually comes down to a single decision made early: whether they are brought into the design of the change as a contributor, or simply informed of it once it is finished. Given a real stake in shaping the solution, and visible credit for the expertise they bring to it, this is often the person who ends up defending the new process most fiercely once the consultant is gone, because they had a genuine hand in building it.
A handover that addresses environmental inertia but ignores enterprise inertia will watch incentives quietly pull the operation backward. One that solves both but ignores the personal dimension may still lose its best people in the process.
Genuine stickiness requires accounting for all four, because inertia only needs to win in one dimension to eventually win in all of them.
There is a fifth complication worth naming, though it is a different kind of problem entirely. Fulfilment and warehouse operations function as ecosystems, not as sets of independent processes, a theme explored in the first article in this series. A fix can hold exactly as designed and still fail to deliver its value, because the process being fixed was never as isolated as it looked. A redesigned pick process that speeds up picking can create a bottleneck at packing that did not exist before. This is not reversion, and it will not be solved by anything addressing inertia directly. It is solved by asking, before a change goes live, what else touches this process and what currently depends on it working the way it does today.
An ongoing force, not a one-time push
Combatting inertia requires an ongoing force, not a one-time push. A launch event, a training week, a well-produced set of documents, these can all contribute to move an operation but they do not hold it there. Holding requires something that keeps applying force after the engagement ends.
James Clear's Four Laws of Behaviour Change, from Atomic Habits, offer a useful lens here: make it obvious, make it attractive, make it easy and make it satisfying. Making a new process obvious means the environment itself points people toward it, the direct antidote to environmental inertia. Making it attractive means a sponsor has genuinely staked their credibility on it and incentives have actually been realigned, the antidote to enterprise inertia. Making it easy means removing friction and not leaving the old way sitting there as the path of least resistance. Making it satisfying is where a feedback loop earns its place, because a team that can see the new way is genuinely working has a reason to keep choosing it.
The force that inertia actually tests. A senior stakeholder who signs off on a solution and then treats the problem as solved was never genuinely sponsoring it, only procuring it. Genuine sponsorship shows up in the sponsor asking about the change at the three-month mark, not just the go-live date, and demonstrating the willingness to spend some of their own authority correcting a team that has started drifting back. Without a sponsor who keeps showing up after go-live, enterprise and cultural inertia operate unchallenged.
What turns a one-time push into an ongoing force. The most powerful and most underused loop available to most operations is customer feedback on the product of the change, not internal throughput metrics, but the signal that comes directly from the person the process exists to serve. This loop also solves a timing problem: most operational improvements take months to deliver their full value, and in that window a partially reverted process and a genuinely successful one look almost identical on the metrics leadership is watching. By the time the gap becomes undeniable, drift has often had months to entrench itself. A feedback loop built around a closer, more immediate signal surfaces the problem while correction is still cheap.
Closes the loop on where genuine ownership comes from. A process designed entirely by a consultant and delivered to the floor as a finished artefact starts with a credibility deficit, because the people expected to run it were never asked what would work. Structured change management exists precisely to close that gap, and skipping it because the technical solution is strong is one of the more common and avoidable mistakes in operational improvement.
The mechanics that make all of the above durable. Training that survives turnover is designed for repetition, not a single delivery, and teaches the reasoning behind a process, not just its steps. Documentation that survives contact with a real operation is written for the edge cases, not just the ideal flow, and has an owner whose job includes keeping it current. And ownership, the pillar that matters most, means someone inside the business understands not just what a process does but why it was designed that way. The clearest test of genuine ownership is whether that person can explain the reasoning to someone else, unprompted. If they can only describe what to do, ownership has not transferred. If they can describe why, it has.
Designed from the first week, not the last
The uncomfortable implication is that a shorter engagement ending without genuine ownership transfer is often worse than a longer one that builds it properly, even if the shorter one looks more efficient on a project timeline. Speed to completion is not the same as speed to durability, and the six-month check is a far more honest measure of success than the go-live date.
The operations that hold their improvements are the ones with a sponsor who stayed invested past sign-off, a workforce that helped shape the solution, including the people whose expertise once made them indispensable to the old way of working, an owner who understands the reasoning behind the process, and a live feedback loop that catches drift while correction is still cheap. That combination is what actually stands between an improvement and the inertia working quietly against it. It is worth designing for from the first week of an engagement, not the last.