Key Takeaways
Treat adoption, not installation, as the real finish line
A working system is not enough. Change management means preparing people to use new tools and processes successfully, not simply delivering the technology. At Acme Corporation, employees liked their customized purchasing system, UBuy, while IT struggled to maintain it. Leaders chose a standardized replacement, We Shop, but staff worried about usability, job security, and losing department-specific processes. A technically flawless launch could still fail if employees avoided the new system or kept using the old one alongside it.
Define success in behavior. Specify what people must do differently, how consistently they must do it, and which organizational benefits should follow. The book cites an IBM survey in which 41 percent of projects fully met their objectives; a separate McKinsey survey found only about one-third of executives said their organizations achieved change.
The distinction between delivery and adoption resembles the difference between installing a tool and changing a working routine. In implementation science, an intervention is not effective merely because it is available; its reach, use, and fit with local practice matter too. That makes the book's focus on lasting behavior more useful than a narrow go-live checklist. Still, adoption alone is not proof of success: people can use a system that fails to deliver promised savings or creates hidden work. Strong measures therefore connect usage to both user behavior and business outcomes.
Show people the problem, then make the new path easier
Facts rarely move people alone. The book pairs rational explanation with emotional engagement and practical ease. Kotter's See-Feel-Change approach asks leaders to make the need for change vivid, rather than relying only on analysis. In one purchasing example, executives had heard arguments for standardization without acting. A display of more than 400 identical rubber gloves, bought by different departments at prices ranging from under $5 to over $15, made waste tangible and helped win their support.
Address three kinds of resistance. The Rider, Elephant, Path metaphor from Chip and Dan Heath distinguishes a person's reasoning, emotional motivation, and the practical route to action. A change plan should answer sensible questions, acknowledge feelings such as fear or frustration, and remove obstacles through usable tools, clear instructions, and accessible help.
This approach aligns with behavioral science's attention to salience and friction. A concrete object can make an abstract cost legible, while a simpler process can reduce the effort required to act. The glove display is persuasive because it turns an invisible system problem into a shared, inspectable fact. Yet emotional vividness can also distort judgment: a striking example may overstate how typical the problem is. A sound case combines memorable evidence with representative data, then gives people a workable alternative rather than merely making them feel urgency.
Coordinate the five change disciplines instead of building another silo
Change work has five connected parts. The book organizes the work around sponsorship, organization design, stakeholder management, communication, and training. Sponsorship provides the foundation; the other areas help people understand, prepare for, and sustain the change. The parts continually affect one another. For example, stakeholder concerns shape training and messages, while system design can alter job duties and the support structure.
Integrate with the project. Project management, functional, technical, and change teams need to plan together. The change team can bring user concerns into design, while technical and business teams provide the information needed for useful training. Partnerships with Corporate Communications, Corporate Training, Human Resources, and other project teams can prevent conflicting messages, scheduling clashes, and incompatible changes.
The five-part structure is valuable because it treats organizational readiness as coordinated work, not a communications campaign added near launch. It also resembles systems thinking: a change in one component can produce effects elsewhere, including workload shifts, new skills needs, and competing demands on attention. The practical test is whether the teams share decisions and timelines, not whether they appear together on an organization chart. Integration has a cost, though. Without clear ownership, collaboration can become a meeting-heavy substitute for decisions. Each partnership needs a defined purpose and a route for resolving conflicts.
Put visible sponsors and trusted peers inside one change network
Support must travel through the organization. A change network combines four groups: the steering committee, executive sponsors, change agents, and super users. Senior leaders set direction, remove barriers, and provide resources. Executive sponsors advocate for the project and visibly demonstrate support. Change agents are respected colleagues who explain the change locally and relay concerns. Super users take on deeper roles, such as testing, training, and front-line help.
Make support two-way. These people should not only promote the project. They also report rumors, questions, and practical obstacles back to the team. For a complex rollout, the book suggests roughly one super user for every 15 to 20 end users. Their managers need to approve the time commitment, and the project should provide information, training, tools, and recognition.
A network of credible peers can bridge the gap between formal authority and everyday practice. Organizational research often distinguishes a policy's official sponsor from the people whose local judgment colleagues actually trust. The book's design uses both: executives supply authority, while peers supply proximity and context. The arrangement works only if representatives can speak honestly. If they are expected to repeat cheerful messages while suppressing genuine problems, they lose credibility and the feedback channel collapses. Recognition matters, but so do protected time and permission to raise inconvenient evidence.
Map each group's influence, fears, and practical changes separately
Do not treat employees as one audience. A stakeholder analysis records who is affected, how much influence they have, their current support, and the support the project needs from them. A skeptical executive with authority to block a decision deserves different attention from an occasional user with little influence. Reassess support as the project develops because people's views can change when they understand the consequences.
Trace the impact. A Change Impact Assessment examines tools and technology, processes, and people. Compare current and future tools against intended benefits, then compare current and future process steps. Finally, identify who gains an opportunity or faces a barrier. Acme's new software alone would not reduce wasteful spending; redesigned purchasing processes were needed to realize that benefit.
This analysis combines power mapping with the practical detail of workflow analysis. That combination prevents a common planning error: assuming a technical feature automatically creates a business result. The distinction between a tool change and a process change is particularly useful because it exposes where training, staffing, or policy decisions are actually required. The method also needs care. Influence ratings can reflect managerial perceptions more than real informal power, and individual stakeholder records can be sensitive. Good analysis should be updated through conversation and handled discreetly, rather than treated as a permanent label attached to a person.
Turn every announcement into a conversation with a clear next step
Sending is not the same as communicating. An email can be delivered without being read, understood, or acted on. Two-way communication gives people a way to ask questions and gives the project team evidence about whether the message landed. Feedback can come through surveys, focus groups, project email, suggestion boxes, change agents, or super users. If the project solicits feedback, it should acknowledge it and explain what happened to it.
Match the channel to the audience. Push channels, such as targeted email, give the sender more control over timing. Pull channels, such as an online resource library, let people find information when needed. Use both when a message is important. A manager cascade can add personal context, but messages can mutate or stop as they pass from manager to manager, so provide a reliable backup channel.
The chapter's key insight is that communication is a feedback loop, not a broadcast event. This fits the logic of control systems: a message without information about its effect cannot be reliably adjusted. The advice to choose channels based on audience and urgency also resists the fantasy of a universal medium. A nuance is that read receipts and survey counts measure contact more readily than comprehension or trust. Better checks ask people to explain what changes for them or demonstrate the requested action. Feedback channels are useful only when participants believe that speaking up is safe and worthwhile.
Choose training by task, audience, and risk, not convenience
Different people need different instruction. Blended learning combines methods such as instructor-led classes, e-learning, coaching, knowledge transfer, on-the-job practice, and self-service guides. Use interactive instruction for difficult topics or questions, e-learning for dispersed groups and simpler material, and coaching for individual needs. Super users need deeper preparation than ordinary users, while executives may benefit from brief, tailored coaching. Training should explain both where to click and how the new process changes the work.
Allow real development time. The book estimates about 200 development hours per hour of fully interactive e-learning, 100 for semi-interactive e-learning, and 40 for classroom instruction. These estimates include review work, and calendar delays can stretch the process further. Pilot courses, gather feedback, and improve them before broad delivery.
The method is strongest when training is treated as performance preparation rather than a compliance checkbox. Matching instruction to task complexity and risk is consistent with instructional design: people need practice, feedback, and a realistic chance to perform the work. The development ratios also counter the temptation to assume polished digital learning is cheap or instant. They are planning estimates, not universal laws; existing materials, software, reviewer availability, and production standards can change the effort substantially. A useful plan tests its assumptions early and protects time for revision, rather than treating the estimate as a fixed promise.
Design future jobs and support roles before the project team disbands
Technology changes the organization around it. New processes can shift duties, create skill gaps, or change the number of people needed. In Acme, department purchasers who once entered requests could move toward vendor negotiation and purchasing analysis, while employees would submit their own requests. IT staff who specialized in UBuy might need training for new roles. Work with Human Resources to review job descriptions, staffing needs, retraining options, and any employment consequences.
Build the operating structure early. Identify the run team that will maintain the system and processes, and the support organization that will answer user questions. Bring these people into the project while design and testing are under way. Waiting until launch leaves too little time to transfer knowledge and can create an unsupported system just as the project team exits.
This is a welcome expansion of change management beyond communication and training. In sociotechnical systems, a tool's performance depends on the surrounding roles, incentives, skills, and decision rights. Planning those elements early can reduce both employee anxiety and operational fragility. It also makes a project confront distributional questions: efficiency gains may create better work for some people while eliminating tasks or positions for others. The book appropriately directs employment decisions toward HR, but consultation should be substantive. A future-state chart is not enough unless affected employees understand what is changing and have credible paths to prepare.
Keep measuring adoption after go-live, when productivity may dip twice
Expect a messy transition. The book describes a pre-launch performance dip, when attention shifts toward the coming change, and a post-launch dip, as people learn new tools and routines. These disruptions vary by group and do not follow a neat timetable. Its change curve tracks movement from unawareness through awareness, understanding, exploration, and adoption, while recognizing that people can pause or move backward. At adoption, users should be able to perform the work confidently and regard the new process as normal.
Keep support alive. Continue communication, training, troubleshooting, and feedback after launch. Sustainable training includes named owners, an update process, and accessible materials for new staff, missed sessions, and occasional tasks. Track whether the change is actually helping, and revise the plan when evidence shows a barrier or a benefit is not materializing.
The two-dip model is a useful reminder that change consumes attention before it changes workflow, then requires learning after implementation. It can help leaders interpret a short-term decline without mistaking every setback for failure. The curve should remain a diagnostic, not a forecast: its stages are not a validated schedule, and productivity changes can also reflect staffing, seasonality, or unrelated pressures. Pairing the model with local evidence is essential. Measures such as error rates, workarounds, support requests, and user confidence can show whether the transition is improving, while business outcomes test whether the original case for change was sound.
Analysis
Carr's practical argument is that IT change is a coordinated organizational intervention whose success depends on whether people can and do work differently. Its distinctive contribution is less a single theory than a connected toolkit: define outcomes, understand affected groups, build sponsorship, assess changes to technology, process, and roles, communicate in both directions, train for actual tasks, and prepare the operating organization to take over. The Acme story gives those tools a concrete spine. Employees value UBuy for reasons leaders initially overlook, so the challenge is not merely persuading them that We Shop is better. It is reconciling standardization with local work, addressing threats to identity and employment, and ensuring the new arrangements function after project staff leave.
The book's strongest design choice is its insistence on integration. Communication depends on stakeholder analysis; training depends on process design; process changes can alter job descriptions and staffing; sponsors can remove barriers none of the other activities can resolve. This resembles a systems approach in which failure often emerges at the seams between components. The planning templates help teams make those seams visible and assign responsibility.
Its evidence is persuasive as a practical warning, but less suited to precise prediction. The cited project-success surveys establish that change efforts often disappoint, yet the summary statistics do not prove that a specific management activity causes success. Likewise, the change curve and performance dips are useful planning metaphors, not universal laws of human response. Organizations should use them to ask better questions, then test assumptions with local feedback and operational measures.
The deeper lesson is that adoption cannot be commanded into existence. Leaders can set direction and expectations, but employees need credible reasons, meaningful involvement, sufficient capability, and workable systems. A robust change plan therefore balances advocacy with listening, and metrics with lived experience. The goal is not to make every person enthusiastic. It is to make the future state understandable, practicable, supportable, and demonstrably better at the work the organization actually needs done.
Review Summary
Practical Change Management for IT Projects receives mixed reviews. Readers appreciate its practical approach, detailed methodology, and focus on communication. The book is praised for its step-by-step guide and use of a case study. However, some criticize its optimistic tone and lack of agile methodologies. Reviewers note it's best suited for large organizations and managers rather than engineers. While it provides a broad overview of change management, some feel it may not offer ready solutions for complex scenarios. Overall, readers find it a valuable resource for understanding change management basics.
Glossary
Five Pillars of Change
Five connected change disciplinesCarr's organizing model for change management activities: sponsorship, organization design, stakeholder management, communication, and training. The components are interdependent and should be planned as a coordinated program. Sponsorship is described as the foundation that supports the work, while the other areas help people prepare for and adopt the change.
Change network
People who support and relay changeA group that connects the project with employees across the organization. It includes senior decision makers, executive sponsors, change agents, and super users. Members communicate the purpose and practical details of the change, encourage participation, surface concerns, and pass feedback between end users and the project team.
Change Impact Assessment (CIA)
Analysis of change effectsA structured review of how a project affects tools and technology, business processes, and people. It compares current and future states, identifies positive opportunities and negative barriers, and links those effects to actions such as communication, training, staffing, or process support. The assessment helps distinguish a technology change from a process change.
Change curve
Stages of change adoptionA conceptual model for tracking how people respond as a change becomes known and incorporated into daily work. Carr describes stages progressing from unawareness to awareness, understanding, exploration, and adoption. Individuals and groups may move at different speeds, pause, or regress, so the curve is a guide for planning support rather than a fixed timetable.
Valley of Despair
Temporary performance declineThe productivity decline that can accompany organizational change. Carr describes a dip before implementation, as attention is diverted by the coming change, and another afterward, while people learn new tools and routines. The severity and duration can differ across groups. Clear communication, appropriate training, and active support can help reduce the disruption.
Blended learning
Combined training methodsA training approach that combines delivery methods to suit the subject, audience, and setting. Options in the book include instructor-led training, e-learning, coaching, knowledge transfer, on-the-job training, and self-service materials. The aim is to choose forms that fit the complexity of the task, participants' needs, geographic distribution, and available resources.
See - Feel - Change
Make the need emotionally vividA change framework associated in the book with John Kotter. Instead of relying only on analysis and rational arguments, it makes the problem or opportunity visible in a way that helps people feel why action matters. The example of displaying differently priced, identical gloves makes purchasing waste concrete to decision makers.
Rider, Elephant, Path
Reason, emotion, and practical routeA framework associated in the book with Chip and Dan Heath. The Rider represents deliberate reasoning, the Elephant represents emotional motivation, and the Path represents the practical conditions for action. Applied to change, it suggests addressing people's questions, engaging their feelings, and removing obstacles that make the new behavior difficult.
Manager cascade
Message passed through managersA communication method in which a leader gives a message to managers, who pass it to their direct reports, continuing through the organization. It can make messages more personal and allow questions in local teams. Because details can be distorted or omitted as they travel, Carr recommends pairing a cascade with a dependable secondary channel.
Download PDF
Download EPUB
.epub digital book format is ideal for reading ebooks on phones, tablets, and e-readers.