Collective Immunity to Change Map: A Worked Guide

A collective Immunity to Change map helps a group examine a recurring pattern it wants to change but continues to sustain. The map connects a shared improvement goal, contrary behavior, possible protective commitments and assumptions the group can investigate. Its value lies in making a working explanation available for examination, rather than declaring that the organization has one hidden motive.

Organizational mapping is especially useful when the pattern crosses functions. One team may seek a quick answer while another withholds provisional information because it has previously been treated as a promise. Each response can reinforce the difficulty the other group is trying to avoid.

This guide develops that interface through a complete fictional map, a bounded five-request test, mixed evidence and an actual decision not to expand yet. The related collective team-mapping guide considers choosing between a shared pattern and individual work; here the emphasis is the organizational loop and how to test its different explanations.

Understand what makes the map collective

Robert Kegan and Lisa Laskow Lahey’s Immunity to Change applies the framework to groups as well as individuals. Their collective approach concerns patterns that people enact together. It does not require every member to disclose the same personal fear or the group to be assigned a developmental stage.

The developers’ current Minds at Work explanation describes four connected elements: a meaningful improvement goal, behavior working against it, the protection that behavior serves, and assumptions that make the protection feel necessary. The protective commitment is more than a worry. It describes what people are trying to prevent or preserve through the behavior.

For example, “We worry about looking unresponsive” names a concern. “We are committed to avoiding an appearance that we cannot handle a request” describes a possible protection. “If we report uncertainty, management will always treat us as incompetent” is an assumption connecting the concern to action. Participants need to recognize or revise those entries; the coach cannot discover them on their behalf by assertion.

A collective pattern may include different protections. Accounts staff could be protecting credibility with management while delivery staff protect their work from unauthorized commitments. A useful map preserves the distinction and asks how the behaviors interact. It does not average the concerns into a single organizational personality.

Check for a real shared goal and a real operational barrier

Before mapping, ask what each function means by improvement. “Respond faster” may mean sending any answer to one person and giving a dependable date to another. A shared phrase can hide incompatible instructions.

Have participants complete two original prompts: “An improvement would be observable when…” and “The work must still protect…”. Compare their answers before asking about hidden commitments. If one group is rewarded for immediate dates and another is held responsible for dates it did not approve, the incentive conflict belongs in the record.

Some difficulties need a straightforward operational repair. A backup without access to a calendar cannot check it by changing a belief. People may also be correctly anticipating consequences: provisional information may really have been misused. Do not design an experiment on the assumption that every concern is irrational.

The developers’ Harvard Graduate School of Education interview explains the gap between wanting a change and sustaining patterns that prevent it. Treat that account as a framework for inquiry. It is not evidence that every missed response in an organization has an adaptive psychological cause.

Worked example: map an organizational loop

All people, systems, requests and results in this case are fictional. A training provider is preparing possible dates for custom workshops. Bilal leads accounts, Chen coordinates faculty availability, Ana manages delivery operations and Dov leads the proposal team. Mira, the managing director, controls the internal pilot and expansion decision.

The four participants voluntarily agree to a 90-minute facilitated meeting on 6 November 2026 and a 60-minute review on 20 November. Their coach, Rafi, works with the group. Mira receives attendance and the shared map and review wording that participants approve, not everyone’s private notes. Work information stays in the provider’s authorized systems; no customer correspondence is copied into a public coaching space.

1. Start with a pattern people can check

The group reviews three recent internal proposal exchanges. Accounts inserted possible workshop dates before faculty confirmation. Faculty then delayed responding until they could give a firm answer. Proposal staff escalated the apparent silence to Mira, who asked for a date urgently. The pressure made accounts more likely to supply a date early in the next exchange.

These observations are a small set of work examples, not proof of everyone’s intentions. Ana rejects the first suggested explanation that faculty are protecting their status. She points to a period when the designated backup lacked calendar access. Chen also says a provisional date had previously lost its qualification when copied onward. Rafi records both concerns rather than treating disagreement with his suggestion as evidence of resistance.

Bilal identifies a further pressure: the current internal reporting target favors fast proposal turnaround. Mira confirms that target exists. She accepts that the pilot will assess its agreed process rather than penalize someone for declining to invent a date. The target outside the pilot has not been permanently changed.

2. Complete the four-column working map

The participants agree to the following language for this work pattern. Their acceptance means they consider it useful enough to investigate, not that it explains every proposal or every individual.

The group lists backup access separately as a factual dependency. It is not placed in the assumption column as “We believe access is missing.” They can inspect the actual permissions. They also retain the reporting-target conflict as an organizational condition, rather than disguising it as a private fear.

This makes the map collective in a practical sense: each function can see its contribution to a sequence and the conditions affecting it. No one is required to agree that they personally hold both assumptions. No national, clinical or developmental label follows from the entries.

Facilitate the map without forcing agreement or disclosure

A first meeting needs enough structure to prevent the map becoming a contest over who is at fault. The case’s 90-minute agenda is an original facilitation choice: ten minutes for purpose and information boundaries, 15 for the shared goal, 20 for concrete behavior and contrary examples, 20 for protections and assumptions, 20 for a possible test, and five for checking the actual record.

Participants may write before speaking, correct the facilitator, decline a personal question or say that the proposed explanation does not fit. A leader can name a work behavior without requiring intimate disclosure. In this case Mira acknowledges that her urgent requests contribute to the loop; she does not ask the four participants to prove trust by sharing personal histories.

Where people face retaliation or other unacceptable conduct, use the appropriate organizational route. A group coach should not promise that a mapping session makes disclosure safe or anonymous. The HSE’s Relationships standard identifies organizational arrangements for raising and responding to unacceptable behavior. A learning exercise cannot replace those arrangements.

If a common goal is genuinely absent, stop short of a collective test. Record the incompatible purposes and the decision needed. It may be more useful for an authorized leader to settle a trade-off than for the coach to produce a map everyone signs without believing.

Design a test that can answer a limited question

Kegan and Lahey describe tests that are safe and modest, attempted soon and approached as research. The purpose is to learn about an assumption, not merely to demonstrate improved performance. A good result for the work goal and useful information about a belief are related but different.

Mira authorizes five routine internal requests received during 9–13 November. Chen is the primary faculty owner. Rosa, the faculty assistant, accepts backup duty for this pilot before it starts; Ana maintains the request log. Urgent incidents, already-promised customer dates and binding contracts are excluded. Nothing in the test authorizes an external date commitment. The request and response record stays in the existing company workspace.

For each request, accounts supplies the workshop question and known constraints. The designated faculty owner or backup sends a status by 17:00 on the next working day. Monday to Friday are working days, and all times use Europe/Amsterdam. The status must say whether a date is confirmed, what remains unknown and who owns the next check. If it includes a planning range, that range must remain explicitly provisional when copied onward.

The completed test design

Notice the limit of the design. It examines internal management reactions and preservation of qualifications. It cannot tell the group whether customers would accept a later answer or whether the sales target is commercially appropriate. Those would require different evidence and authorization.

Follow the five actual requests

The table records the completed fictional pilot. Each of the five original replies includes an unconfirmed status and an explicitly provisional planning range. A next-working-day deadline means a Friday request is due on Monday, not Saturday. The receipt window and final response window are therefore different.

Keep the access defect in the result

P3’s primary owner is absent. The backup can open the request but cannot read the designated availability calendar. Ana checks that defect and escalates it. Chen eventually supplies the availability information for the late reply, which arrives after the Thursday deadline. Calling the result a lack of commitment would obscure an actual system condition.

With Mira’s approval, Ravi, the company’s access owner, gives the designated backup the limited calendar permission on 17 November. On 19 November Ana and Rosa complete a synthetic check: Rosa opens the intended calendar and returns a status from a made-up request. That verifies this narrow access repair. It is not another eligible live request or evidence that all future replies will be timely.

Retain the range that lost its qualification

For P4, Chen’s original response states that a possible week is a planning range, not a confirmed workshop date, and names the remaining faculty check. Bilal copies only the possible week into an internal message to Mira and Dov. Chen sees the omission and asks that the affected request be paused.

Bilal checks the original and sends this correction to the same recipients on 13 November: “My previous message omitted the qualification. No workshop date is confirmed. The possible week is a planning range only, pending Chen’s faculty check. Please do not use it as a commitment.” Mira and Dov acknowledge the correction before the request proceeds further internally.

The correction matters, but it does not make the original transfer accurate. Faculty’s concern has received supporting evidence in this case. The group needs a better information boundary rather than pressure on Chen to abandon the concern.

Listen to the actual management response

Dov challenges the unconfirmed P2 status: “This is incomplete. What can the proposal team use now?” Bilal initially experiences the reply as evidence that he has looked incompetent. He checks rather than recording the interpretation as fact.

Dov explains that he needs the next decision owner and an expected update, not an invented date. The original response had named the owner but omitted the update time. Chen agrees to post the next status in the existing company thread by Thursday 12 November at 15:00. Bilal sends that accepted update time, and Dov confirms that this answers his immediate process question. Chen posts the update at 14:30 on 12 November: faculty availability is still unconfirmed and no workshop date may be promised. The request remains unconfirmed as a workshop booking.

This is mixed information for accounts. A clear unconfirmed status did receive criticism, and one response was incomplete. The exchange did not include a judgment that accounts was incompetent. That limited observation weakens the necessity of the original prediction in this instance without proving management will always respond constructively.

Reconcile goal performance and assumption learning

At the 20 November review, Ana brings the five original response records and their onward messages. Four of five were timely. Four of five retained the required qualification in the initial onward transfer. Only three met both conditions on the first pass: P1, P2 and P5. P3 was late; P4 was timely but lost its qualification. This does not mean the other three were fully useful on arrival: P2 still needed an agreed update time.

Those are three different measures. Five eventual responses do not mean five timely, accurate responses. Later corrections must be recorded as corrections rather than used to erase the initial defect. The synthetic access check remains outside the five-request denominator.

The group also revises its explanation. Accounts no longer treats every request for clarification as a necessary verdict of incompetence, but retains the actual criticism and incomplete P2 response. Faculty sees four qualifications preserved and one stripped. The broad “inevitably disappears” claim does not fit all five events, while the underlying risk clearly remains real.

Chen declines the suggestion to extend provisional dates into customer messages. The other participants accept that this is a separate risk decision, not evidence that Chen failed the mapping exercise. Ana proposes improving the internal response template to require an agreed update time. Bilal proposes preserving the complete qualification whenever a range is forwarded.

All four approve a short review for Mira: five internal requests, four timely, four initially qualified onward, three meeting both conditions, one later qualification correction, and a backup-access repair checked only synthetically. Mira receives and acknowledges it. She does not authorize expansion because the revised transfer method and the wider target conflict still need decisions. No future customer test or permanent incentive change has occurred by this cutoff.

Use a practical record for the next organizational map

The next record should connect the framework to observable work, not collect personal disclosures because a worksheet has space for them.

1. Write the shared improvement and its boundary. State what better work would look like and which decisions remain outside the group.
2. Describe the recurring sequence. Include what each function does and how others respond. Add examples that do not fit the pattern.
3. Keep worries and protections distinct. Ask what people fear and what they may be trying to prevent. Let participants reject the suggested explanation.
4. Record more than one assumption when needed. Connect each prediction to the relevant behavior rather than forcing a single hidden cause.
5. Separate dependencies. Check permissions, capacity, incentives and authority as facts or unresolved decisions in their own right.
6. Design a bounded research action. Name the eligible events, actual owners, information route, stop condition and evidence that could change the interpretation.
7. Review original and repaired results. Preserve late replies, missing qualifications and adverse reactions. State what remains unknown before deciding whether to continue.

A map can be revised, narrowed or set aside. If the shared goal is no longer meaningful, or the obstacle is now an explicit resource decision, repeating the hidden-commitment discussion may add little. The executive team-coaching framework helps place that decision within the wider engagement.

Keep the collective record and private work separate in Coachful

Coachful is our #1 recommendation for organizing the original preparation, shared learning material and chosen follow-up for this work. Its group, resource and client-plan features support the coaching process; they do not diagnose an organizational immunity or authorize a business change.

1. Prepare the learning resource. A short Course can explain the exercise and its boundaries before the meeting. Use your own authorized material and examples; a course template or copied private link does not grant access to every participant.
2. Choose the shared audience. If a group community holds the agreed map and review, inspect actual membership and staff permissions. Do not post raw individual maps merely because participants share a group.
3. Protect unshared reflection. Unshared journal and takeaway pages remain private to the client. Where enabled, Share with coach deliberately shares a chosen page. Invite an agreed work summary rather than requiring sensitive personal words to be shared.
4. Review actions and outputs accurately. Keep coach-team review notes separate from the approved sponsor output. Check recap audience, the workspace Approve before sending setting and session-type overrides. A checked task does not establish another person’s agreement or correct a company calendar permission.

In the case, the shared map and the participants’ approved short review are the agreed organizational outputs. Rafi does not export private preparation to Mira or represent the group’s hopeful interpretation as a tested organizational result.

Frequently asked questions

Must everyone share the same competing commitment?

No. Different functions may protect different things while reinforcing a common pattern. Preserve those differences and test the relevant predictions separately. Do not assign an averaged fear to every participant.

Can the map replace individual coaching?

A collective map examines shared work patterns. People may have private goals they do not want to explore with colleagues. Do not require their individual maps as the admission price for participating in the team work.

What if the assumption turns out to be partly true?

Keep the supporting evidence. In this case a qualification really was lost, so the group retained that risk and corrected the actual recipient. Learning may mean strengthening a boundary rather than deciding the concern was mistaken.

How long does overcoming an organizational immunity take?

There is no universal completion time established by this example. A meeting can produce a working map, but later events and decisions determine what is learned. This case ends with a limited review and blocked expansion, not a claim that an immunity has been dismantled.

When should we stop mapping and make a management decision?

When the obstacle is a known authority, resource or incentive conflict, take it to the person who can resolve it. The map may explain the pattern around that obstacle, but it does not supply missing permission or decide the organization’s trade-off.