Watch the recording of the webinar:
Listen to the recording via our podcast:
View the Slides:
H1: No More "Flavor of the Month": How to Deliver Sustainable Change That Sticks
Page title (60 chars): No More Flavor of the Month: Sustainable Change
Meta description (155 chars): Lynn Kelley on why 70% of organizational change fails — and the change questions framework she used to reach 96% sustainment at Union Pacific.
Focus keyphrase: sustainable change continuous improvement
Slug suggestion: no-more-flavor-of-the-month-sustainable-change (use existing recording slug if it ranks; otherwise migrate and 301 the registration page)
Tags: Change Management, Sustainability, Continuous Improvement, Lean Leadership, Change Questions
A KaiNexus webinar with D. Lynn Kelley, Ph.D., former SVP of Supply Chain & Continuous Improvement at Union Pacific Railroad
[Video embed]
[Slide deck embed or link]
[Podcast audio embed]
Free digital workbook at change-questions.net →
Most organizations have lived through the cycle. A new improvement initiative launches. Leadership announces it with energy. Teams are trained. Posters go up. Six months later, the program is quietly dying. Twelve months later, no one mentions it anymore. Eighteen months later, a new initiative launches, and the people who were burned the first time keep their heads down. The phrase "flavor of the month" gets used, sometimes out loud, often just thought.
The cost of that cycle isn't just the wasted resources of the failed initiative. The bigger cost is the damage left behind. People who participated and saw the change collapse become harder to engage next time. The next initiative starts with a higher resistance threshold than the last one did. Eventually, the organization develops a kind of immune response to change, and even genuinely good initiatives can't get traction.
Lynn Kelley has spent her career studying this problem from the inside. She held executive leadership roles at Textron, including Vice President of Operational Excellence, where she rolled out change in 32 countries. She served as Senior Vice President of Supply Chain and Continuous Improvement at Union Pacific Railroad, where she led improvement work across 42,000 employees in the operating department. She holds a Ph.D. in evaluation and research and taught graduate-level statistics before moving into industry. After retiring from Union Pacific in 2018, she co-authored Change Questions with John Shook, distilling decades of experience into a practical framework anyone can use.
The session walks through that framework — eleven questions organized into five categories — drawn from research and tested across hundreds of change initiatives. The Union Pacific case study runs through the book and surfaces throughout the talk. The result Lynn shares at the close is the headline: change initiatives at Union Pacific were sustained at a 96% rate using the framework, against an industry baseline that runs between 50% and 70% failure. The session is the explanation for how that gap closed.
What follows is the substance of the session, organized so the page is useful whether you watched it or are landing here from search.
The numbers Lynn opens with are worth sitting with for a moment.
She asks the audience to estimate the percentage of people who keep their New Year's resolutions. The answers cluster in the single digits. The actual figure: about 9%. By the first week of January, 23% of people have already abandoned their resolutions. By the end of January, 43% have given up. Personal change — change someone has chosen, that affects only themselves, with no political dynamics involved — is hard.
Then she asks about organizational change. The estimates run higher: 50%, 70%, sometimes 80%. The research consistently puts the failure rate between 50% and 70% depending on how it's measured. The Harvard Business Review article that surfaced this framing in 2000, "Cracking the Code of Change," gave the higher number. McKinsey's research has reached similar conclusions. Even taking the most generous interpretation of the data, half of organizational change initiatives fail to sustain.
The implication that matters most: every failed initiative leaves damaged goods behind. People who lived through it become harder to engage next time. The next initiative starts with a higher resistance threshold. The organization develops a flavor-of-the-month immune response. The cost compounds.
Lynn's framing throughout: even if your organization is operating at the 50% failure rate (the optimistic end of the data), that's still too high to be sustainable, because each failure makes the next one more likely.
A specific point Lynn raises that changes how change practitioners should monitor their initiatives.
She asks the audience: how soon after launch can you tell whether a change is going to succeed or fail? The audience guesses range from one month to six months to "immediately based on management reaction."
The research-based answer: within the first month. There are enough early signals — engagement levels, leader behavior, the quality of questions being asked, whether the work is actually happening — that the trajectory becomes visible quickly. The implication is that change practitioners shouldn't wait six months to know whether something is working. They should be looking actively in the first 30 days for the signals, ready to pivot when something is off.
This connects to the strategic speed concept later in the session. Most implementation plans are timeline-based — do this by this date, then this by this date, then this by this date. Better implementation plans build in checkpoints from the first weeks, not just at the end. The earlier you detect a problem, the cheaper it is to fix.
Lynn's framework didn't come from a single research project or theoretical framework. It came from years of failing at change and studying why each failure happened.
She traces the origin to her time at Textron starting in 1998. She had been hired after leading TQM rollouts at Automotive Fastener plants in Michigan, including a program that won the Michigan Quality Leadership Award. Textron acquired the company and asked her to move to Paris and replicate the work in their French and German plants — five or six in each country.
The German plants succeeded, partly because Mercedes and BMW were pressuring their suppliers to adopt Toyota-style practices. The French plants didn't. Her success rate was 50%. Her boss noticed.
That failure started a research habit. After each unsuccessful change, Lynn would dig into peer-reviewed research on what had gone wrong. Over time, the research surfaced a list of questions she should have been asking before launching. The list grew from three to four to eventually eleven. The questions clustered into five categories. By the time she ran change at Union Pacific, she was using the questions on every initiative.
John Shook, who Lynn served on a board with, was working on his own change book during COVID. They started as writing buddies, sharing chapters monthly. Two months in, they realized they were writing the same book. The overlap between Lynn's questions (drawn from research and corporate experience) and John's questions (drawn from his global Lean network experience) was about 85-90%. What he had, she didn't, and vice versa. They merged the lists and co-authored Change Questions.
The framework is available two ways. The book is about 200 pages with extensive case material from Union Pacific. The free digital workbook at change-questions.net is a 40-page fillable PDF that walks through the eleven questions with sample agendas, deliverables, and team action items. Both lead to the same framework.
The questions cluster into five categories that, together, cover what most change initiatives need to address.
Purpose. What problem is this change trying to solve, and what's the value the organization expects to get? The category exists because most change initiatives skip past purpose to action, and people downstream are left guessing why they're being asked to change.
Designing and doing the work. What's the actual work changing? How will it be tested? How will success be detected early? This is where pilots, change agents, and the early signals of success or failure live.
Engaging and developing the employees. Who needs to be engaged, how, and with what kind of support? Communication, recognition, training, and the human side of change all sit here.
Management system and leadership. How will leaders visibly support the change? What management behaviors need to shift? How will managers reinforce the new way of working through their daily routines?
Culture. What's the existing culture, how can it support the change, and what cultural friction needs to be mitigated? Culture is the catch-all category because culture permeates the other four — but it also gets explicit attention so it doesn't get assumed away.
The framework is flexible. Lynn recommends starting with purpose because it grounds everything else. Beyond that, the questions can be answered in any order that makes sense for the specific change. Not every question applies to every initiative. The point is to think through them deliberately rather than launching into action and discovering the gaps later.
Lynn opens the framework with a study most change practitioners don't have on the tip of their tongue.
Three-year-olds ask "why" about eight times a day. Some of them, according to research, ask 30 times a day. The need to understand why isn't a phase. It's an innate human pattern that persists into adulthood and into the workplace. PWC research found that Millennials in particular have the strongest stated need to understand the purpose of their work — but the underlying pattern applies across generations.
The implication for change: if people need to understand why even for their day-to-day work, the need is much greater when they're being asked to change. Asking someone to do something differently — to learn something new, to give up a familiar pattern, to take time away from work they already feel stretched on — is asking them to absorb cost. The cost is more bearable when the purpose is clear.
Lynn's purpose statement format is simple. Two parts: what you're going to initiate or implement, and why. A robust purpose statement includes both. The workbook walks through writing one and refining it. People who already know the answer can write it in a minute. People who can't write it quickly are exactly the people who need this question most — because if they can't articulate the purpose to their own team, they can't expect anyone downstream to absorb it.
One of the strongest moments in the session is the story Lynn uses to introduce the change curve.
She had rolled out a change at Textron to a group of 250 people she knew well — a group with whom she had built trust over years. The change was small, the CEO wanted it, and Lynn assumed she could skip the change questions. She held a meeting with the leaders, said the CEO wanted this done, asked for input, got muted agreement, and rolled out the change by email.
She got slammed.
That weekend she went to Martha's Vineyard with a friend. They were walking along a bay when a group of 10 or 11 swimmers came along the beach with inner tubes. The swimmers walked the entire arc of the bay to the far end. Then one person ran and jumped into the water and was carried across the bay by a riptide that ran horizontally rather than out to sea. Another jumped in. Then four or five floated in on inner tubes and let the current carry them. Then one woman waded up to her knees and waded back. The last few walked all the way back around the bay to meet their friends.
Lynn realized she had just watched the change curve play out in front of her.
In any group, roughly 20% will accept change readily. Roughly 60% will be neutral — not actively resisting, not actively pushing forward, just floating with the current. Roughly 20% will resist initially. The implication: when you launch a change to everyone simultaneously, the loudest voices are the resisters. Those voices pull the neutral middle in their direction. By the time you realize what's happened, the change is in trouble.
The countermeasure she added to her framework that weekend: how can I pilot this change with people who are already open to it? Every change, even ones that seem hard to pilot, can usually be tested in some form — an experiment, a simulation, a soft launch with a willing group. Two reasons. First, pilots surface bugs the practitioner needs to fix before broader rollout — and asking people to figure out the bugs while also accepting the change is asking too much. Second, pilots produce successes that can be publicized, generating the FOMO effect that pulls the neutral middle toward adoption rather than away from it.
This connects to a quote from Malcolm Gladwell's The Tipping Point that Lynn lifts during the session. The neutral middle is the tipping point. Whoever reaches them first — the early adopters or the resisters — determines where the change ends up.
A related change question Lynn surfaces: how can we recognize success?
Because failure or success is detectable within the first month, the practitioner can be actively looking for early wins to publicize. When a pilot produces results, the CEO or senior leader talks about it. The team that achieved it presents at a forum. A podcast episode features them. An article goes out internally.
The mechanism is FOMO. The neutral middle hears that another team got results, got resources, got attention from leadership. They start asking when they get to participate. The dynamic flips from push to pull. Instead of the practitioner trying to convince teams to adopt, teams are asking to be next.
This works because the recognition isn't manufactured. The pilot actually produced results. The team actually did the work. The recognition is real, and the FOMO is a downstream consequence of the recognition being real. Recognition without underlying results would produce a different dynamic — cynicism rather than pull.
A McKinsey study Lynn references makes a distinction most change initiatives fail to make.
Of the change initiatives that fail (which the study put at 70%), 33% fail specifically because management behaviors did not support the change. The wording matters. The study didn't say management endorsement. It said management behaviors. Most managers, when asked, will endorse the change. They'll tell their boss they support it. They'll say the right things in town halls. But their daily behaviors — what they review, what they reward, what they ignore, what they skip — communicate the actual organizational priority.
The implication for the change questions: don't ask managers if they support the change. Ask managers how they will visibly show their teams they support the change. The shift in question changes the psychological dynamic. Asking for endorsement produces compliance. Asking for visible behavior produces commitment, because the manager is now mentally rehearsing the specific behaviors they'll perform — and people accept what they help to create.
The framework also asks: do you support this change, and if not, why? And what would it take for you to support it? Naming the resistance creates space for it to be addressed rather than hidden. Managers who are quietly skeptical produce more damage than managers who openly disagree, because the open disagreement can be engaged.
A study Lynn references from Harvard Business Review of more than 300 companies implementing change.
Two patterns emerged. The first, operational speed, was timeline-based: do this by this date, then this, then this. Marching orders with a Gantt chart. Moderately successful.
The second, strategic speed, used the same overall plan but built checkpoints throughout the implementation. The plan wasn't mandatory in the marching-orders sense. It was a hypothesis that would be tested at each checkpoint. If the change was producing the expected results, the team continued. If not, they tweaked or pivoted. The implementation was treated as a series of experiments rather than a march.
The follow-up data: companies using strategic speed had 52% better results than companies using operational speed.
The mechanism is the early-detection point from earlier in the session. Most change failures are detectable in the first month. A timeline-based plan doesn't pause to look. A checkpoint-based plan pauses at week two, week four, week six — actively asking whether the early signals are showing up. When the signals are off, the team adjusts before the rollout has propagated through the rest of the organization.
Lynn's recommendation: build at least one checkpoint into the first month of any change initiative. Measure the leading indicators, not just the lagging financial outcomes. If the engagement isn't where it should be, if the behavior changes aren't visible, if the early adopters aren't getting traction — pivot before the failure becomes structural.
A reframe Lynn surfaces near the end of the session, attributed to John Shook.
When she initially wrote up the change questions framework, she described it as "putting the time in ahead of time" — doing the upfront work of purpose, pilot design, leader engagement, and so on before launching the change. John pushed back. The framing was wrong. The upfront work isn't preparation for the change. It is the change. The diagnostic, the design, the leader engagement, the pilot — all of it is part of what the change consists of, not setup that precedes the real work.
The reframe matters because it changes how teams allocate effort. Teams that view the upfront work as preparation skip it under deadline pressure. Teams that view the upfront work as the actual change protect that time the way they'd protect any other critical work. The shift in framing is small. The downstream effect on whether the framework actually gets used is large.
Worth surfacing because Lynn's framework assumes it without always naming it.
Several change questions only work in environments where people will tell the truth. The question "do you support this change, and if not, why?" only produces useful data if managers feel safe saying no. The question "what's getting in the way of adoption?" only produces useful data if the people asked feel safe naming real obstacles rather than diplomatic ones. The early-detection mechanism — looking for signals of failure in the first month — only works if the people who see the signals are willing to surface them rather than hide them.
The pilot mechanism is also a psychological safety move. Asking people to figure out a half-baked change creates risk for them. Asking them to participate in a tested, debugged change reduces that risk. Lynn's framing — "we should respect them enough that we don't give them something full of problems they shouldn't have to figure out" — names the principle without using the safety language directly. The principle is the same one underneath psychological safety: treating people as people whose time, judgment, and engagement matter.
The connection runs deeper in the management behavior question. Asking managers how they'll visibly support the change — instead of just whether they support it — is a move that respects their autonomy to design their own commitment. People accept what they help to create. The question creates ownership rather than compliance, which is one of the practical mechanisms by which psychological safety produces engagement.
CI practitioners using the change questions framework benefit from recognizing that the questions assume a baseline level of organizational safety. In environments where safety is low, the questions still help — they surface what people are actually thinking once the right space is created — but the practitioner may need to invest in safety building before the questions produce reliable answers.
The closing slide of the session is the headline.
Across hundreds of change initiatives at Union Pacific, the sustainment rate when Lynn left in 2018 was 96%. At Textron before that, the rate was 90%. Both numbers are dramatically higher than the 30%-50% baseline rate that the broader research consistently shows.
The measurement was rigorous. Sustainment was defined as six consecutive months at or above the KPI target the change was intended to move. Initiatives had a year to reach that bar. If the metrics started slipping in the early months, the team would be pulled back in to figure out what was happening before the slide became permanent. Most of the metrics were electronic — train movements, stops, operational data — which removed the possibility of fudging.
A related set of practices reinforced sustainment. Process owners were renamed "sustainers" to make their ongoing accountability explicit. Projects didn't "close" — they "transitioned" to the sustainer. The language signaled that the work continued past the launch milestone, which is where most change initiatives quietly stop maintaining their gains.
The results are evidence that the failure rate isn't inevitable. Organizations that work the change questions deliberately, build sustainment into the design from the start, and use rigorous measurement to verify the gains are holding can dramatically outperform the industry baseline.
A few specific things the platform does that connect to the framework.
KaiNexus tracks improvement work through implementation and beyond, which gives change leaders the visibility into early signals that strategic speed depends on. Checkpoints become structural rather than dependent on individual practitioner discipline.
The platform supports the recognition mechanism Lynn describes. When a pilot produces results, the work is visible to the rest of the organization automatically. The FOMO effect that pulls the neutral middle toward adoption gets amplified by the platform rather than depending on manual communication.
It also supports the sustainment measurement Lynn used at Union Pacific. KPIs can be tracked over time with the consecutive-month threshold built in. Process owners — sustainers, in Lynn's language — have explicit ongoing visibility into whether the gains are holding. When metrics start to drift, the platform surfaces the drift early enough to pull the team back in.
If your organization is producing good initial results from change initiatives but watching them quietly erode in the months after launch, the gap is usually the infrastructure that makes sustainment visible and accountable. That's the gap KaiNexus is built to close.
D. Lynn Kelley, Ph.D. retired from Union Pacific Railroad in 2018, where she served as Senior Vice President of Supply Chain and Continuous Improvement and as the executive co-owner of the company's innovation program. Before Union Pacific, she was Vice President of Operational Excellence, an officer, and a member of the executive leadership team at Textron. She holds a Ph.D. in evaluation and research and taught undergraduate and graduate statistics. Earlier in her career, she served as Executive Vice President and Chief Operating Officer of Doctors Hospital in Detroit. She now serves as a senior advisor to BBH Capital Partners, supporting investment evaluation, transaction execution, and post-investment oversight of portfolio companies. She is the co-author with John Shook of Change Questions: A Playbook for Effective and Lasting Organizational Change.
Why do most organizational change initiatives fail?
The research consistently puts the failure rate between 50% and 70% depending on how it's measured. The most common drivers are insufficient visible leadership commitment, poor communication, management behaviors that don't reinforce the change, and rollout patterns that overweight the resistant 20% by launching to everyone simultaneously. Each failed initiative also leaves damaged goods behind — people who participated and saw the change collapse become harder to engage next time, which makes the next initiative harder than the last.
What are the change questions and where did they come from?
The change questions are an eleven-question framework developed independently by Lynn Kelley and John Shook over decades of change leadership experience and research, then merged when they realized they were writing the same book. The questions are organized into five categories: purpose, designing and doing the work, engaging and developing employees, management system and leadership, and culture. The framework is available in the book Change Questions and as a free fillable digital workbook at change-questions.net.
How quickly can you tell whether a change initiative is going to succeed or fail?
Within the first month. Research shows there are enough early signals — engagement levels, leader behavior, the quality of questions being asked, whether the work is actually happening — that the trajectory becomes visible quickly. The implication for change leaders is to build checkpoints into the first 30 days rather than waiting six months to discover that something has gone wrong.
What's the difference between operational speed and strategic speed?
Operational speed is timeline-based change implementation: do this by this date, then this, then this. Strategic speed uses the same plan but builds in checkpoints throughout, treating implementation as a series of experiments. At each checkpoint, the team verifies whether the change is producing the expected results. If not, they tweak or pivot before the rollout propagates further. Research from Harvard Business Review found that companies using strategic speed had 52% better results than companies using operational speed.
Why is launching a change to everyone simultaneously usually a mistake?
Because of the change curve. In any group, roughly 20% will accept change readily, 60% will be neutral, and 20% will resist initially. When you launch to everyone simultaneously, the resisters have the loudest voices and pull the neutral middle in their direction before the change has time to demonstrate value. Piloting with people who are already open to change lets you debug the change, produce visible early wins, and create FOMO that pulls the neutral middle toward adoption rather than away from it.
How is "engagement" different from "endorsement" for senior leaders?
Endorsement is what most leaders give a change initiative when asked: they say the right things, they sign off on resources, they show up at kickoff meetings. Engagement is what their daily behaviors actually communicate. Of change initiatives that fail, McKinsey research found that 33% fail specifically because management behaviors didn't support the change — even though management endorsement was usually present. The change questions framework asks managers not whether they support the change but how they will visibly show their teams that they do, which produces commitment rather than compliance.
What sustainment results did Union Pacific achieve using this framework?
When Lynn left Union Pacific in 2018, the sustainment rate across hundreds of change initiatives was 96%. Sustainment was measured rigorously: six consecutive months at or above the KPI target. If metrics started slipping in the early months, the team was pulled back in to address it before the slide became permanent. The 96% rate compares against a baseline industry failure rate of 50%-70%, which is evidence that the failure rate isn't inevitable when the upfront work is done deliberately.
.svg)
Copyright © 2026
Privacy Policy