Summary. Head office decides a change, and the store is where it either happens or quietly does not. Most change programs are designed for the deciding and resourced for the announcing, with almost nothing left for the part that actually determines the outcome. This paper is about the last hundred feet: why the floor triages rather than resists, what it costs when a change only half lands, and how to tell the difference before the quarter closes.

The floor is already at capacity

Any change arriving in a store lands on people who are not idle. Retail Systems Research, in fieldwork run from December 2024 through January 2025, found that 45 percent of store employees spend too much time trying to find answers to customer service questions, that 42 percent of retailers report associates losing time to technology support and maintenance, and that 38 percent report time lost to administrative tasks.¹

Read that as the operating conditions for every initiative you launch. The change is not competing with spare capacity. It is competing with the tasks already filling the shift, and with a customer who is about to ask something.

This is why the standard failure mode is not resistance. It is triage. An associate who cannot do everything does the things that are visible, checked, and asked about by their manager. Everything else becomes optional in practice regardless of how it was labeled in the announcement.

Triage is worth taking seriously because it is rational behavior, not a discipline failure. The associate is optimizing against the signals they actually receive, and if the only signal attached to your change is an email, they have correctly ranked it below the thing their manager will ask about at four o’clock.

Four things that decide whether a change lands

Whether the store understands why. Not the strategy, the store-level why. An associate asked to change how they greet customers needs to know what it is meant to achieve, because they will encounter the situation the instruction did not anticipate within the first hour and will have to improvise from principle.

Whether the manager was consulted before launch. A store or district manager who first hears about a change at the same moment as their team cannot sponsor it. They can only relay it, and relayed instructions are followed at relay-level enthusiasm.

Whether it replaced something. Change programs almost always add. If nothing was removed, the associate resolves the conflict by dropping whatever is least likely to be noticed, and head office never learns which.

Whether anyone checked. Confirmation is what separates a change from an announcement. Not a completion percentage at fleet level, but knowing which stores actually adopted it and which are still doing it the old way.

“Our people are professionals, they do what they are asked”

This objection is usually raised sincerely, and by someone who has been in stores and knows the teams. It deserves an answer rather than a raised eyebrow.

Two things are true at once. Store teams are overwhelmingly willing, and the change still does not land uniformly. The reason is arithmetic rather than attitude.

Willingness does not create hours. If the shift is full before your change arrives, willingness determines which task gets dropped, not whether one does.

The variance is the evidence. Take any recent change and look at adoption by store. If the same instruction, delivered the same way to the same kind of team, produced a wide spread, the difference is not professionalism. It is what happened locally around the instruction, which is the thing you can design.

Nobody reports a quiet non-adoption. A store that has dropped a step is not concealing it. They are prioritizing, and there is no field in any system where they would enter that decision. Absence of complaints is not evidence of adoption.

The concession worth making: a change that store teams believe in will land through almost any process, badly designed or not. That is exactly why the store-level why belongs at the front of the briefing rather than in an appendix.

The sequencing most retailers get wrong

The usual order is decide, announce, train, then measure whether it worked. The order that survives contact with a sales floor is close to the reverse.

1. Decide what you will measure before you write the content, because that decision determines what the content has to teach.
2. Brief the field first, giving store and district managers enough lead time to ask questions and to raise the practical objection you did not think of.
3. Say what stops. Explicitly, by name. If nothing stops, explain what should be deprioritized when the day is full.
4. Deliver in the flow of work, short enough to complete between customers rather than requiring a session.
5. Confirm per store, then per person, and follow up the exceptions rather than re-broadcasting to everyone.
6. Come back in three weeks with the same message in a different form, because a single exposure does not change a habit.

Step three is the one that gets cut, and it is the one the floor notices. A change announced without a corresponding subtraction tells store teams that head office does not know what their day contains, and that impression is expensive to reverse.

What partial adoption costs

Change programs are usually evaluated as though they either worked or did not. In practice they land in some stores and not others, and the cost of that is calculable.

1. Take the result the change was supposed to produce and measure it in your top adopting quartile against your bottom quartile, on comparable stores.
2. Treat the gap as the value of adoption, not as a store quality difference, provided the stores are genuinely comparable on format, footfall and staffing.
3. Apply that gap to the stores below the line. That is the unrealized return on a program you have already paid to design, produce and announce.
4. Compare it to the cost of the program itself. In most cases the unrealized portion exceeds the entire delivery budget, which is the argument for spending more on the last hundred feet rather than less.

The useful property of this calculation is that it needs no new instrumentation. You already have the outcome metric and you already know which stores adopted. Putting the two together is the analysis that almost never gets run.

The pilot that tells you nothing

Most changes are piloted, and most pilots are designed in a way that guarantees a favorable answer.

The usual pattern picks strong stores, run by managers who volunteered, with a project team available by phone and a visit from someone senior in week one. The change works. It then rolls out to four hundred stores with none of those conditions present, and performs nothing like the pilot.

Three corrections make a pilot informative.

Include stores you expect to struggle. A pilot that only proves the change works where everything else works has answered a question you were not asking.

Withdraw the project team. After the first week, let the stores run it through the normal support route. Whatever breaks then is what will break at scale.

Hold a matched set back. Comparable stores that do not receive the change give you a control, and without one the promotional calendar can explain any result you get. Eight weeks with a held-back group is worth more than six months of before-and-after.

The point of a pilot is not to build confidence for the launch meeting. It is to find the two things about the change that nobody in the room could have predicted, while they are still cheap to fix.

Why the pace of change makes this structural

A retailer used to run a handful of major operational changes a year. Now the product calendar, the promotional calendar, the technology roadmap and the compliance cycle all generate their own changes, and they arrive at the same store on the same Tuesday.

That volume means change management cannot be a project team assembled per initiative. It has to be a standing capability: one route into the store, an agreed way of signaling priority, and a habit of confirming receipt. Retailers who build that find each subsequent change cheaper than the last, because the pipework already exists and the floor already trusts it.

Retailers who do not end up with what most have: several parallel channels, no agreed priority, and an associate deciding for themselves what matters. That governance problem is worth solving before the next initiative.

What to measure

1. Days from decision to the store being able to act. Not to the announcement. To the point where an associate could do the new thing correctly.
2. Adoption spread across stores at week two and week six. If the spread widens between the two, the change is decaying rather than embedding.
3. What was named as stopping, and whether stores agree it stopped. A subtraction that did not actually happen is the most common cause of a change that fades.
4. Manager lead time. Hours between the field briefing and the team briefing, tracked per initiative. It is the cheapest lever in the whole list.
5. Repeat exposure completion. Whether the three-week follow-up reached the people who needed it, rather than the people who already complied.

Where to start

Take the last significant change you rolled out. Ask three stores, chosen by a district manager who will be honest with you, three questions: what were you asked to do, what did you stop doing to make room, and who checked. The gap between the answers and the plan is your actual change management capability.

Confirming a change without turning it into policing is the part most retailers find hardest, and it is mostly a design choice. The calculator will model what closing the adoption gap is worth on your own numbers.

References

1. Retail Systems Research, sponsored by Jumpmind; fieldwork December 2024 through January 2025, reported by Retail Customer Experience. Findings on store employee time spent locating answers, on technology support and on administrative tasks. retailcustomerexperience.com