A change agent lets a small, contained failure happen in a unit that resists the change, then lets the visible collapse make the case that the old way no longer works.
An example
A medical device maker was moving its field service teams to a centralized dispatch model, and the regional service director for the Southeast had argued for keeping local scheduling, which her team's response times supported. In the run-up to a large hospital system's annual equipment audit, her requests for two loaner technicians and access to the new parts-forecasting tool went unanswered, while other regions got both. The audit found eleven overdue preventive maintenance records, and the hospital system escalated to the company's chief commercial officer. At the next operating review the transformation lead said nothing about centralization. The overdue records were on the screen, and the region moved to central dispatch the following quarter. A year later, the director's dated email requests, forwarded to a new head of service, reopened the question of what the region had been given.
Why it works
Once the old way collapses in plain view, it looks worse, because its defenders' claims have been contradicted by a failure everyone watched. Defending it then looks costlier: the defenders must deny the failure, call it an anomaly without evidence, or take the blame. The change agent's restraint is the mechanism, since arguing for the new approach during the failure would expose a motive. The conditions are a failure that can be contained, plausible deniability for the withheld help, a failure clear enough that observers reach the conclusion unprompted, and defenders who cannot prove sabotage.
The thinking behind it
Amos Tversky Tversky and Kahneman's Judgment under Uncertainty: Heuristics and Biases (1974) describes availability: judging likelihood by how easily instances come to mind. A failure at a major customer's audit is vivid and recent, so it becomes the instance everyone recalls when the old method comes up. Dated records of the help that was withheld supply a competing instance, which is what lets the failure be reread.
Lee Ross Ross's The Intuitive Psychologist and His Shortcomings: Distortions in the Attribution Process (1977) names the fundamental attribution error: observers explain behavior by personal traits and underestimate the situation. People who saw the region fail its audit explain it by the team and its method, and miss the technicians and tools it was denied. Putting the withheld support on the timeline makes the situation visible.
Warning signs
Support requests from one unit go unanswered in the run-up to its visible failure, while peer units receive help. The unit is the one most resistant to the change. The change agent says little about the new approach during the period. The failure lands just before a forum where it will be seen, such as an audit, a board meeting, or a customer review.
How to defend yourself
Document every request for help and its answer before the failure lands: dates, recipients, and what peer units received. The tactic depends on the failure looking self-inflicted. Once the change agent's fingerprints show on the abandonment, the failure indicts the change agent and leaves the old way standing.
- If support is being withheld, say so in writing to the person who could release it, and copy someone senior.
- After a failure, put the withheld support on the timeline, so the cause is traced to the abandonment.
- Ask for your methods to be judged on the periods when you were resourced like your peers.
How a leader shuts it down
Track support requests and response times across units going through a change, and look for one unit getting less. Review any failure in a resistant unit for what it was given beforehand. Compare old and new methods with equal resources, and treat a convenient failure as a reason to look harder.