How to Spread Continuous Improvement Across Your Organization

Most organizations don't have a problem starting continuous improvement. They have a problem spreading it.

A few departments take to it quickly. They generate ideas, implement changes, build momentum, and produce visible results. Other departments stall. Some never start. The leadership team looks at the heat map of activity across the organization and sees a few green pockets, a lot of yellow, and stubborn red zones that haven't moved despite the same training, the same tools, and the same expectations.

After a year or two, the conclusion often becomes "we tried Lean" or "we tried Kaizen" -- when in truth the methodology was deployed in a few places but never actually spread to most of the organization. The improvement program technically existed. It just didn't take.

Spread is the work that turns improvement from a project into a culture. It's also the work most improvement programs underinvest in, because spread is slow, requires sustained leadership attention, and doesn't produce visible artifacts the way the initial pilot did. The early enthusiasm gets the program started. Spread is what keeps it going beyond the first generation of champions.

This webinar is a conversation between Mark Graban and Greg Jacobson, MD on what spread actually requires -- the three elements that have to be present together, the leadership behaviors that sustain it, and the patterns that distinguish organizations that scale improvement from organizations that talk about it.

Mark Graban is a Senior Advisor to KaiNexus and the author of Lean Hospitals and co-author of Healthcare Kaizen. Greg Jacobson, MD is co-founder and CEO of KaiNexus. The discussion draws on hundreds of conversations with organizations across healthcare, manufacturing, and professional services about why spread succeeds or fails.

What "spread" actually means

The Japanese term is yokoten -- the lateral spread of learning, practices, and successful improvements across an organization. It's distinct from rollout in an important way. Rollout implies top-down deployment: leadership decides on a practice, mandates it, and pushes it through the organization. Spread is more organic. It implies that a practice or idea developed in one place is being adopted, adapted, and improved upon elsewhere because people in those other places see value in it.

The distinction matters because rollouts often produce compliance without engagement. People do the thing because they were told to. Spread, done well, produces both adoption and ownership. People do the thing because they understand why it works and have made it their own.

What gets spread isn't a single artifact. It can be a methodology (how problem-solving happens in the organization), a leadership behavior (how managers respond to staff ideas), a specific implemented improvement (a process change that worked in one unit and could work in others), or all three layered together. The work of spread operates at all of these levels simultaneously.

How ideas spread, generally

Before getting to the specific tactics for continuous improvement, Mark and Greg take a step back to look at how ideas spread in general. The reference is Atul Gawande's 2013 New Yorker article "Slow Ideas," which examines why some innovations spread quickly and others take generations to take hold.

Gawande's example pair is striking. Surgical anesthesia, discovered in 1846, spread to the capitals of Europe by February of that year and to most regions of the world by June 1847. This was the era of carrier pigeon and telegram, and yet within months, the practice had reached operating rooms worldwide.

Surgical antisepsis, discovered by Joseph Lister and others in 1867, took decades to spread. Many surgeons resisted it actively, and even today, hand hygiene compliance in hospitals remains an ongoing improvement target -- nearly 160 years after the underlying science was established.

The difference Gawande identifies is visible benefit. Anesthesia produced an immediate, dramatic, visible result -- patients stopped thrashing on the operating table, surgeons could do their work, the benefit was undeniable in the moment of use. Antisepsis produced an invisible benefit -- patients who didn't get infected, weeks after the surgery, often in someone else's care. The cause-effect relationship was abstract. Worse, early antiseptic methods caused painful stinging on surgeons' hands. In the short term, adopting antisepsis made the surgeon's life worse while improving outcomes for patients who weren't yet sick.

The relevance to spreading continuous improvement is direct. Big visible wins spread fast. Subtle systemic improvements that require sustained discipline and don't produce immediate dramatic results spread slowly, if at all. Most of the work of continuous improvement falls in the second category. The implication: spread requires deliberate effort precisely because the underlying ideas don't spread on their own.

Gawande's recommendations for accelerating spread of slow ideas include one-on-one mentoring (not just classroom training), multiple touch points (he suggests seven on average, not one), and key memorable messages tied to a clear purpose. Each maps directly to how continuous improvement actually spreads in organizations.

The Frank Gilbreth example is in the same vein. In 1915, the industrial engineer Frank Gilbreth presented findings to the American Medical Association showing that surgeons spent more time searching for instruments than performing surgery. He proposed what he called a surgical caddy -- the idea of a dedicated person handing the surgeon instruments on cue. It took the AMA 15 years to formally bless the idea. By 1930, what is now standard surgical practice was finally endorsed.

The pattern of healthcare being slow to adopt operational ideas isn't ancient history. Don Berwick's seminal 1989 paper on applying continuous improvement principles to healthcare -- which is the founding document of much of the modern Lean Healthcare movement -- is now nearly four decades old. The ideas have spread to some organizations beautifully. They remain entirely absent in others.

The three things spread requires

The structural argument in the session is that successful spread depends on three elements operating together: methodology, leadership, and technology. When organizations struggle to spread improvement, one of these three is missing or weak. When they succeed, all three are present and reinforcing each other.

This is not a tools-and-techniques framework. It's an operational systems framework. The point is that each element supports the others and that no one element is sufficient on its own.

Methodology is the structured approach to improvement that the organization uses. The argument is not that any particular methodology (Lean, Six Sigma, Toyota Production System, Model for Improvement) is the right one. It's that whichever methodology the organization picks needs to share three characteristics. It needs to be consistent across the organization, disciplined in its practice, and simple enough that people who aren't improvement specialists can engage with it.

Consistency doesn't mean identical. Every department has different work, different staffing, different needs. A consistent methodology means the same underlying principles and the same recognizable structure -- the same way of asking questions, the same way of organizing improvement boards, the same vocabulary for what an improvement looks like. The implementation can be tailored. The thinking is shared.

Discipline means continuous, not sporadic. Most "continuous improvement" programs are actually periodic improvement programs. Activity spikes during events. Activity ebbs between events. The improvement is real but the continuity isn't. A disciplined methodology operates on a sustained cadence -- daily huddles, weekly reviews, monthly reflections -- not as a series of bursts.

Simplicity is the constraint most organizations get wrong. The reflex of improvement specialists is to teach everything they know -- the full Lean toolkit, the complete DMAIC methodology, the entire Toyota approach. This produces resistance, not adoption. Frontline staff don't need to become improvement experts. They need a simple enough framework that they can participate without having to learn a discipline. The improvement specialists exist to coach and support, not to require everyone else to become one.

Leadership is the second leg. Specifically, three leadership behaviors at every level of the organization: commitment to continuous improvement, communication of that commitment on a sustained and frequent basis, and accountability that holds the practice in place. These don't have to consume 40 hours a week of leadership time. They can be done in five or ten minutes a day, but they have to happen consistently. The phrase "the lifestyle, not the fad diet" captures the difference between organizations where this becomes the operating norm and organizations where it's a campaign that eventually ends.

Technology is the third leg. The argument here isn't about the specific platform. It's about what technology enables: visibility across the organization, standardization of the methodology, collaboration across silos, and aggregated measurement of impact. Spreadsheets, SharePoint, bulletin boards, and email newsletters all attempt some of these functions and all run into limits as the organization scales. The visibility breaks down because the artifacts are local. The standardization fails because each implementation drifts. The collaboration is difficult because the silos remain. The measurement is impossible because nobody has a complete view.

What separates organizations that succeed at spread from organizations that don't is rarely just one of these three. It's usually all three operating together.

The "go slow to go fast" of spread

A counterintuitive principle Mark and Greg keep returning to: spread is faster long-term when it starts slow.

The reflex is to roll out improvement programs aggressively. Train everyone. Deploy bulletin boards in 150 departments. Launch the technology platform in every unit simultaneously. Hold a kickoff event. The intent is to demonstrate seriousness and to capture early enthusiasm.

The result is usually predictable. The organization can't support 150 departments simultaneously. Coaches are stretched too thin. Managers in 130 of those departments don't get the help they need. The methodology doesn't take. People conclude the program failed and move on to the next initiative.

The alternative is deliberate slow spread. Start with two or three departments. Coach them intensively. Get them genuinely good at the methodology. Use their success to demonstrate that the approach works and to create pull from other departments who want to be next. Add new departments at a rate that the central coaching team can actually support.

One customer Mark references in the session began this way. They started with two departments in December. By April -- four months in -- they were active in 14 departments with four more starting in the next couple of weeks. Their reflection: active coaching is key. The lead improvement specialist wasn't running a one-hour onboarding session and walking away. He was spending two days per department over two weeks, shoulder-to-shoulder with managers and frontline staff, working through real improvements together. That depth of engagement is what produced the spread velocity.

The math is worth noting. Two departments to fourteen in four months sounds slow if you expected to deploy across the organization on day one. It is fast if the comparison is the more common pattern of deploying everywhere on day one, watching most of it fail, and finding yourself back at zero on day 365.

How departments get selected for spread

A practical question that comes up early in any spread effort: which departments go next?

The approach Mark and Greg recommend, and which they've seen work, is to follow pull rather than push. The departments that go next are the ones whose managers want to be next. Managers volunteer. They actively ask "when do we get to do this?" The presence of that pull is the signal that the cultural conditions exist for the methodology to take hold.

This is intrinsic motivation in Daniel Pink's framing -- the motivation that comes from genuine interest and ownership, not from external pressure. Intrinsic motivation is durable. Extrinsic motivation, even when it produces initial activity, fades as soon as the external pressure relents.

The reasonable next question: what happens when you run out of departments with intrinsic motivation? At some point in the spread, you exhaust the volunteers and have to find ways to engage the reluctant. The session offers two answers.

The first is creating pull by making the work of the early departments visible. Monthly forums where teams share their improvements -- both the big Kaizen events and the small daily improvements -- expose other departments to what's possible. People in the audience start thinking about whether something similar could work in their area. Awareness creates curiosity, and curiosity creates pull where there was none before.

The second is leadership messaging that signals priority. When the CEO talks about continuous improvement repeatedly, in formal settings and informal ones, the message gets through. The customer Mark references received top-level state Baldrige recognition. In the CEO's acceptance speech, daily continuous improvement was named as central to ongoing success. That kind of public commitment from the top makes it harder for managers to ignore the work, even ones who weren't initially drawn to it.

The combination of pull from peers and signal from leadership eventually reaches departments that would never have volunteered on their own. They don't become rock stars overnight. They start somewhere. Over time, with coaching, they get better.

Spreading specific improvements, not just methodology

The session distinguishes between two kinds of spread: methodology spread (the work of getting consistent improvement practice across the organization) and idea spread (the work of taking a specific successful improvement from one place and adapting it to others).

The basic hypothesis for idea spread: if a problem existed in one department and was solved, the same problem probably exists in other departments and the solution might help them too.

The example Mark uses is concrete. A procedure room in one unit wasn't stocked with the equipment physicians needed, slowing down patient care. The local improvement put locks on the cabinet doors and added a routine of asking physicians what supplies they wanted available. The problem-solution pair seems mundane until you ask whether it exists elsewhere in the same hospital, or in other hospitals in the same system. It almost certainly does.

The barrier to this kind of spread is that ideas tend to stay where they're generated. Without infrastructure for sharing them, they live in the department that came up with them and never travel. The session names this the "down the well" pattern -- the idea gets implemented locally, the local team takes credit, and nobody outside the silo ever hears about it.

Two related ideas are worth pulling out.

Start small, then spread. The temptation when an idea looks good is to roll it out everywhere at once. The patient call-light caddy example illustrates the failure mode. If the hospital had bought 200 of them and distributed them universally, they'd have learned about the design flaws too late. Instead, they piloted a few designs in a single unit, learned that patients also needed to store glasses and cellphones, iterated, and only then expanded. The improvement that ultimately spread was better than the one they would have rolled out on day one.

Beware "rollout" framing. When ideas are spread by mandate -- "this worked, now everyone has to do it" -- people often feel rolled over rather than included. They become compliant rather than engaged. They stop thinking about the idea and how it might be improved further. The improvement freezes at the version that was mandated.

The ThedaCare model offers an alternative. When an idea works in an alpha unit, it gets shared with a beta unit -- with the explicit invitation to improve it further. The beta unit's improvements feed back to the alpha unit. Both units benefit. The idea continues to evolve rather than ossifying. That model treats spread as a learning process rather than a copying process.

Leadership behaviors that sustain spread

The session closes with the three leadership behaviors that show up consistently in organizations that successfully spread continuous improvement: commitment, communication, and accountability.

These don't sound revolutionary, which is part of the point. The work isn't about discovering new behaviors. It's about doing the basic ones consistently across many levels of the organization for a long time.

Commitment means actually caring about continuous improvement, not just nominally supporting it. The behavioral test: when an improvement initiative competes with another priority for the leader's time, which one wins? Leaders who are committed make time. Leaders who are nominally supportive don't.

Communication means saying things about improvement frequently and consistently. Once a quarter doesn't register. Once a week, with specific examples, with recognition, with the leader's own reflection on what's working and what isn't, does register. Frequency creates the cultural presence that makes improvement feel like the way work gets done, not a side activity.

Accountability is the third and is the one most often softened or skipped. If improvement is genuinely important, then doing the work matters and not doing the work has consequences. The consequences don't have to be punitive. They can be conversational -- the manager who isn't running their huddle gets a coaching conversation, not a write-up. But the conversation has to happen. Without accountability, the message is that improvement is optional, and improvement that's optional doesn't survive the next quarter's competing priorities.

These behaviors operate at every level. The CEO has to model them for the executive team. The executive team has to model them for the directors. The directors have to model them for the managers. The managers have to model them for the frontline staff. The model breaks down anywhere a level skips the work -- the people below that level lose the signal that this matters.

How KaiNexus connects

The technology leg of the three-element framework maps directly to how the platform is designed.

Visibility across the organization -- the ability to see what every department is doing without having to walk to their bulletin board -- is what enables remote coaching, cross-silo idea spread, and aggregate measurement. Without that visibility, even good methodology and good leadership can't scale beyond the people who can physically reach each other.

Standardization of methodology is supported by configurable workflows that ensure the same underlying structure is followed across departments even when local implementation varies. The consistency that the methodology section calls for is operationalized by the platform.

Collaboration across silos is what turns individual improvements into spread improvements. When a team in one department implements something useful, the platform makes it visible and adaptable for other teams who could benefit. The down-the-well problem is structurally addressed by making the well visible to everyone.

Aggregated impact measurement is what lets leadership defend the program to skeptical CFOs. Without aggregate measurement, the improvement work is anecdotal -- you can point to a few stories but you can't show the cumulative effect. With it, you can show the dollars, the hours, the safety incidents prevented, the patient satisfaction gains. That aggregation matters because it's the input that keeps the program funded and prioritized.

The platform doesn't substitute for the methodology or the leadership. It supports both. The argument throughout this session is that all three legs are needed. Skip any one of them and the program eventually stalls.

See KaiNexus in action →

About the presenters

Mark Graban is a Senior Advisor to KaiNexus and an internationally recognized Lean consultant, author, and speaker. He is the author of Lean Hospitals and co-author of Healthcare Kaizen, and has worked extensively in healthcare and other industries to strengthen improvement culture and leadership practices.

Greg Jacobson, MD is co-founder and CEO of KaiNexus. A practicing physician before founding the company, he has focused on the science of organizational culture, habit formation, and the systems that make continuous improvement sustainable at scale.

Frequently Asked Questions

What does "spread" mean in continuous improvement?

Spread is the lateral movement of improvement practices, methodologies, and specific successful changes across the organization -- from pockets of activity to the entire system. The Japanese term is yokoten. Spread is different from rollout, which implies top-down deployment with mandated adoption. Spread is more organic, driven by genuine pull and ongoing adaptation rather than pure compliance. The distinction matters because rollout often produces compliance without engagement, while genuine spread produces both adoption and ownership.

Why do most continuous improvement programs stall after early success?

Because one of three elements is missing or weak: methodology, leadership, or technology. Organizations often invest heavily in one or two of these and underinvest in the third. Methodology without leadership produces strong central CI teams running improvement in isolation from the operations of the business. Leadership without methodology produces enthusiasm without structure. Either of these without technology infrastructure produces local pockets of success that can't scale, because the visibility, standardization, and cross-silo collaboration aren't in place. Successful spread requires all three operating together.

What makes a methodology good for spreading?

Three characteristics: consistency, discipline, and simplicity. Consistency means the same underlying structure across departments, even when local implementation varies. Discipline means continuous and sustained, not periodic bursts of activity. Simplicity means accessible to people who aren't improvement specialists -- the methodology shouldn't require everyone to become an expert in order to participate.

How should organizations choose which departments to spread to first?

By following pull rather than pushing. Start with departments whose managers actively want to engage -- the ones who volunteer, who ask "when do we get to do this?" Intrinsic motivation is durable. Extrinsic motivation fades as soon as the external pressure relents. As the spread progresses and visible success accumulates, departments that weren't initially motivated often become pulled in -- by peer awareness, by leadership messaging, or by seeing concrete examples that work in contexts similar to theirs.

What's the difference between spreading methodology and spreading specific ideas?

Methodology spread is the work of getting consistent improvement practice across the organization -- the same problem-solving approach, the same kind of huddles, the same vocabulary. Idea spread is the work of taking specific successful improvements from one place and adapting them to others. Both are necessary. Methodology spread without idea spread produces consistent practice but doesn't capture the cumulative benefit of learning across the organization. Idea spread without methodology spread produces individual wins that don't compound into a culture.

What are the leadership behaviors that sustain spread?

Three: commitment, communication, and accountability. Commitment means actually caring about continuous improvement when it competes with other priorities for time and attention. Communication means saying things about improvement frequently, with specific examples and recognition, not once a quarter. Accountability means the conversation happens when the work isn't getting done -- not necessarily punitive, but real. These behaviors have to operate at every level of the organization. The model breaks down anywhere a level skips the work, because the people below that level lose the signal that this matters.

See KaiNexus in action →