Keep an Improvement Working After the Trial Ends

An invented sorting trial goes smoothly because its designer stays beside the table, explains every exception, and replaces missing cardboard category labels. The three-tray arrangement looks successful. Yet the exercise has demonstrated a sorting setup plus an attentive designer plus an informal supply service. Removing the extra support creates a different arrangement.

Sustainability begins with that distinction. Ask what made the trial work and which parts would still exist afterward. This article uses only a fictional desktop exercise. It is not permission to implement a change, withdraw a real support, revise a policy, or test a clinical or safety process. Do not use real operational or workforce records.

The trial is not the operating arrangement

IHI’s Implementing Changes guidance distinguishes testing from embedding a lasting change into the way work is done. Supporting documentation, training, and infrastructure may be affected. That general distinction helps explain why a promising prototype does not, by itself, establish readiness for routine use.

Separately, the public summary of IHI’s Sustaining Improvement emphasizes daily management work supported by defined tasks and responsibilities. The fictional failure drill below is this publication’s application of the ownership question, rather than a reproduction of that framework or its organizational cases.

Write the “worked because” list

Return to the imaginary sorting setup. It sorts invented display cards into three labeled trays. List the supports present: the designer answered questions; spare category labels were nearby; every invented card fitted a listed category; the fictional task was not competing with other work. Keep the trays themselves on the list, but do not let them stand for the entire arrangement.

Now mark which supports are part of the proposed design and which are temporary trial assistance. “Someone will remind everyone” is not a complete support description. It leaves the role, trigger, and competing demands unresolved. Similarly, “the supplies will be available” names a desired condition without explaining how the fictional arrangement maintains it.

Run three failures on paper

1. The designer is absent. An invented sorter encounters a card that fits two categories. What part of the proposed arrangement gives an answer, and who is responsible for unresolved questions?
2. A component changes. A new invented display-card type is introduced while the trays still assume the old categories. What would prompt review of the arrangement before the mismatch becomes routine?
3. A support disappears. Spare category labels are no longer available in the fictional setup. Does the arrangement remain understandable, and which role has authority to resolve the gap?

Write the consequence of each failure before writing a solution. If the consequence is repeated sorting, identify where that happens in the invented process. If the consequence is uncertainty, say which decision becomes unclear. Avoid solving every problem with another reminder; reminders also need someone to maintain them.

Specify ownership without assigning real work

For each fictional dependency, name a role rather than a person. The category owner decides classification questions; a materials owner maintains the invented label specification; an approving role decides whether the arrangement may change. One role could cover several functions, but the functions should remain explicit.

This exercise does not assign duties to colleagues. In a real project, responsibility and resources must be agreed through the appropriate local structure. A plan that assumes unapproved work from another team has not solved its sustainability problem. It has transferred the unresolved problem into a sentence.

Build a reason to review, not a promise of permanence

For the paper scenario, choose meaningful review triggers: a component changes, a new exception cannot be resolved within the categories, or the agreed scope no longer fits the task. Triggers are different from vague promises to keep an eye on things. They connect a changed condition to a specific decision.

Include a retirement question. If the imaginary display-card service ends, what would make the sorting setup unnecessary? If a replacement arrangement is approved, who decides when the old one stops applying? Keeping every past improvement indefinitely can create duplicate instructions and maintenance work. Ending an obsolete arrangement can be part of maintaining a useful system.

Finish with the unresolved conditions

The output of this desktop exercise is a short readiness statement: “The invented sorting setup depends on usable categories, replacement labels, and an accountable role for unresolved cases. Those conditions are not yet demonstrated.” That is more useful than declaring the idea sustainable because its first trial went well.

Read the permission-bounded test guide before considering any next step, and the workload-impact guide to examine the continuing maintenance burden. A durable improvement needs a workable surrounding arrangement, not only a successful moment.