The Distance Between Done And Used

Original dramatic voices inspired by historical perspectives; not historical quotations. The companies and workplace cases are fictional composites.

Arena Episode

Morrow Supply sells building materials to local contractors. Its returns desk must decide whether a returned item meets the ordinary refund policy or needs a supervisor's exception. Customers need an answer they can act on before heading back to work.

The company's new assistant reads the order and the clerk's notes, drafts a recommendation, and prepares a customer explanation. Clerks still make ordinary decisions under the existing policy; supervisors approve exceptions. Management wants shorter waits and less repeated explanation. The software is deployed, training is complete, and the usage report looks encouraging. Nobody has yet shown that customers get a dependable answer sooner.

Purpose

Explore what a successful adoption effort should be allowed to remove, including parts of the product it was commissioned to spread.

Players

Case

Ruth, an experienced clerk, makes routine decisions from the policy and uses the assistant afterward to produce the required summary. Eli, a new clerk, starts with the assistant as trained. Both appear on the usage report. Their work takes very different paths.

Stakes

• Customers are waiting for a usable decision while management measures activity in a tool.
• Ruth wants to stop doing duplicate work. Eli wants a way to become competent without pretending he already is.
• The Manager must decide whether to require more use, change the workflow, or withdraw a capability the company has already built.

The Story

After The Customer Leaves

At the returns counter, Ruth opens the assistant beside a completed refund. The customer has already left. Eli watches from the next station. The Manager has brought the Arena visitors to see the new system in use.

Manager: Ruth is one of our most consistent users.

Ruth: Thank you. Can I stop?

Peter Drucker: We have arrived at an unusually efficient awards ceremony.

Manager: Stop using it?

Ruth: Stop telling it what I have already done so it can recommend that I do it.

W. Edwards Deming: Is that how you were trained to use it?

Ruth: No. It is how I get the customer served and the summary filed.

Jane Jacobs: Show us the last return.

Ruth: Unopened box, matching order, within the ordinary policy. I checked the record, approved it, then asked the assistant for the explanation. The record still needs that explanation.

Peter Drucker: So the summary has a job. Does the recommendation?

Manager: It helps people apply the policy consistently.

Mister Rogers: Eli, is that what it helps you do?

Eli: Sometimes. I am still learning what counts as ordinary.

Manager: That is why I want everyone to start there. Otherwise we never give the new process a chance.

Jane Jacobs: Ruth has given it a place. Just not the place on your diagram.

W. Edwards Deming: Which means the usage report cannot tell you which process is being used.

Ruth: It can tell you I was here after the customer wasn't.

The Person Who Followed Instructions

The Manager looks toward the loading desk.

Peter Drucker: Let us see a case where the recommendation mattered.

Eli: This opened box. The customer said the contents were damaged when he collected it. The assistant found the opened-item rule and recommended referring it to a supervisor. I sent the request.

Peter Drucker: And then?

Eli: The customer waited. I could see the request had been sent. I couldn't tell him when anyone would answer.

Ruth: I would have asked the supervisor at the loading desk.

Eli: I didn't know I could. He was helping a contractor collect an order.

Jane Jacobs: There is your working system. A person, a location, a judgment about when to interrupt. The screen has an arrow marked escalation.

Manager: That sounds like an argument for more training.

Mister Rogers: What would you train him to say when the supervisor is busy?

Manager: That a customer is waiting.

Eli: They were both customers.

Peter Drucker: Good. This is a management decision now. Whose wait takes precedence?

Jane Jacobs: Ruth knows how to negotiate it. That knowledge matters.

Eli: Ruth also knows everyone. When I interrupt, I feel as though I am cutting in.

Jane Jacobs: Yes. I was about to recommend following Ruth's path. I should ask whether it is open to you.

Mister Rogers: Did you tell anyone you were stuck?

Eli: I asked whether I had sent it correctly. They said yes.

W. Edwards Deming: You asked about the step. They answered about the step. The customer's problem remained exactly where it was.

Ruth: I thought he needed help with the software.

Eli: I did. But after that I still needed help with the return.

A Better Answer Can Still Be Late

Manager: I don't want to lose what we've built because the first weeks are awkward.

Peter Drucker: Nor should you. But protecting the investment requires finding what it accomplishes. Can the tool finish the ordinary case without drafting a new opinion about the policy?

Manager: We could make the policy checks more direct.

Ruth: Please. I do not need an essay on an unopened box.

Mister Rogers: Eli might need an explanation.

Ruth: He can ask me.

Eli: I already ask you a lot.

Mister Rogers: That is a different cost. A person can be willing to help while the person needing help becomes reluctant to ask.

Peter Drucker: Then a useful explanation is part of the result for a new clerk. I was too quick to treat it as spare packaging.

W. Edwards Deming: Keep that distinction when comparing work. Experienced and new clerks do not have the same starting conditions.

Host: Yesterday's Mastermind recap makes a distinction that may help here: stable rules, AI interpretation, and human authority are different parts of a workflow. A recurring task can become dependable without making every part more intelligent.

Manager: Are you suggesting we take AI out of the returns assistant?

Peter Drucker: Out of the part where a fixed rule is enough. Perhaps.

Jane Jacobs: Keep the interesting question. On an opened box with a disputed history, it could help assemble what happened. That is different from inventing discretion the clerk does not have.

W. Edwards Deming: And a program can route an exception without a model composing the route each time. But the program cannot make a busy supervisor available.

Manager: So the most visible new feature could become the smallest part of the process.

Ruth: It would still get a chair. It wouldn't need the whole counter.

Mister Rogers: Does that feel like failure to you?

Manager: A little. We asked people to learn this. Now we tell them some of it wasn't necessary?

Mister Rogers: You could tell them that using it taught you something. That is a harder sentence if you have already asked them to call it a success.

W. Edwards Deming: Test the simpler route. A tidier training deck will not tell us whether a customer gets an answer.

Who Gets To Be Done?

Manager: The project team dElivered what we asked for. I cannot keep their work open indefinitely.

Peter Drucker: Close their dElivery obligation honestly. Keep the business benefit unproved honestly. Those statements can coexist.

Jane Jacobs: Who owns the space between them?

Manager: Operations. Me.

W. Edwards Deming: With what freedom? May you remove a feature, or only persuade more people to use it?

Manager: I can change the desk procedure. Product changes need the project sponsor.

Peter Drucker: Then you and the sponsor have a decision to make together. Calling you the adoption owner while forbidding subtraction would give you a sales job inside your own company.

Ruth: With us as leads.

Mister Rogers: Reluctant ones, apparently.

Ruth: I have qualified myself as a prospect for fewer screens.

Jane Jacobs: There is an FCG question in this: where is velocity actually needed? Ruth's routine decision is already quick. Eli's wait is somewhere else. Speeding up the same recommendation for both could leave both problems intact.

W. Edwards Deming: And Ruth's quick decision is not automatically correct. We should check the old route too.

Ruth: Fair. Familiar does not mean flawless.

Peter Drucker: I began by asking whether the recommendation had a job. Now I think we need to ask whether this is one job. Ruth needs less duplicate work. Eli needs a rEliable way to learn and decide. The customer needs a sound answer in either case.

Mister Rogers: And Eli needs to be able to say the explanation did not help.

Manager: Without being marked down for failing to adopt it.

W. Edwards Deming: Yes. A compulsory click would make your report easier to complete and your question harder to answer.

Jane Jacobs: I will accept a shared route if newcomers can actually travel it. I will not defend a shortcut that works only when everyone knows your name.

Ruth: Then we should teach it with a return in front of us. Not another tour of the menus.

Eli: Could I do the return, while you watch?

Ruth: Yes. And I will try not to do it with my eyebrows.

Mister Rogers: That may be the advanced course.

One Return In Front Of Us

The Manager opens the record of Eli's delayed return.

Manager: I can settle who covers the exception queue. But what do we test before asking the sponsor to change the product?

W. Edwards Deming: On Monday, take completed returns: ordinary ones and exceptions. Use copies. Have Ruth and Eli each work through examples with the current assistant and with a simpler policy-check route.

Eli: Won't the second way be easier because I have just seen the answer?

W. Edwards Deming: It may. Alternate which route comes first, and record familiarity and help received. We are looking for where work gets stuck. This small exercise cannot prove a general advantage.

Peter Drucker: Follow each attempt to a justified decision and an understandable customer explanation. Starting a request is not finishing a return. Count review, correction, help and waiting, not just the time the assistant is thinking.

Mister Rogers: Tell them what the exercise is for before they begin. No individual league table. Eli must be able to stop and say, “I don't know what to do next.” That is a useful finding.

Jane Jacobs: Include the awkward moment when the supervisor is occupied. Name a backup and a visible waiting state. If nobody can answer, the clerk needs an honest message and the existing supervisor route, not permission to improvise a refund.

Manager: I own the desk procedure and coverage. The sponsor owns the proposed software simplification. Ruth and Eli both attempt the examples during scheduled work. Then Ruth helps Eli practice where he gets stuck. No live refunds from the exercise.

W. Edwards Deming: Before comparing results, you and the policy owner agree what counts as a correct disposition and acceptable explanation. If an example is disputed, label it unresolved. Do not treat the old answer as tRuth merely because it happened first.

Peter Drucker: At Friday's review, decide among keeping the assistant as it is, narrowing it to useful explanation and evidence assembly, or pausing its recommendation step. Bring the sponsor the attempts, not an adoption slogan.

Manager: What would make us stop immediately?

W. Edwards Deming: Any test that touches a live refund. Any recommendation that exceeds the clerk's authority stops that attempt and requires policy review. If the simpler route loses required evidence or misclassifies a return, it has not earned replacement either.

Jane Jacobs: Repair the failed step and repeat the example before resuming that route. Keep the existing authorized service process for customers while you learn.

Mister Rogers: What if Eli needs longer because he is learning?

Peter Drucker: We budget learning time. We do not call it waste, and we do not hide it. If it produces later independence, that may be worth the investment.

W. Edwards Deming: “If” matters. A good session with Ruth beside him cannot establish independence. After the initial review, he needs another agreed, supervised attempt with unfamiliar examples. Further live use needs a separate operating decision.

Manager: Then Friday can authorize the next test. It cannot declare the benefit realized.

Peter Drucker: Correct. The decision can be timely without the conclusion being premature.

The Report Changes

Eli asks Ruth to leave the next example on his screen. This time she steps to the side. The Manager goes to arrange coverage at the loading desk. The proposed test has not proved a benefit. It has given each of them something different to do.

Manager: The usage count may go down if we do this well.

W. Edwards Deming: It may. Customer waiting may go down, stay the same, or reveal another bottleneck. We will have a better question either way.

Ruth: Can the report say I helped remove a step?

Manager: It can say we are testing whether that step should exist.

Ruth: You are getting cautious.

Manager: I have been sitting with Edwards.

W. Edwards Deming: Standing. There aren't enough chairs at this counter.

Jane Jacobs: The assistant has one.

Mister Rogers: Perhaps it can share.

Field Note From The Arena

Morrow's usage report counted Ruth's after-the-fact paperwork and Eli's attempt to make a decision as the same success. Their needs were different, and neither click count established a customer benefit.

Velocity To Value, an FCG operating philosophy, directs attention to the path from an idea to a useful consequence. Here it exposes a limit in adoption work: the owner of the benefit must be able to propose subtraction, including less AI. Innovation Management supports the choice to keep, change, or stop a capability on evidence rather than on its dElivery history.

On Monday, compare actual decision paths using completed examples, including new and experienced users. On Friday, choose the next bounded change; do not confuse that choice with proof of durable value. Preserve correctness, authority and the customer explanation while measuring the whole effort.

Carry this question: Can your adoption effort succeed by making the product smaller?