This one isn't a single-presenter teaching session. Greg Jacobson and Mark Graban decided to experiment with the format — partly because 2020 had pushed them into more panel discussions and conversations than usual, and partly because the standard 45-minute slide deck had started to feel like the default rather than the choice.
What they ran instead was a conversation in four parts. A round of audience polling through live word clouds and rating sliders to find out who was in the room and what state continuous improvement was in across their organizations. Greg's weekly "muse" — the short opening monologue he does at the KaiNexus team meeting — applied to the November 2020 moment of the pandemic. A quick three-question trivia quiz about KaiNexus history. And then the bulk of the hour spent in Ask Us Anything mode, taking questions submitted in advance and live, ranging from how to start a CI culture to how KaiNexus eats its own dog food to what process improvements the pandemic accelerated.
The substantive material is woven through the conversational structure rather than organized into a teaching arc. This page pulls out the threads worth keeping — the recurring themes about leadership, buy-in, recognition, getting started, and the honest picture of how the KaiNexus team uses continuous improvement on itself.
Greg Jacobson, MD is the CEO and co-founder of KaiNexus. He had the original idea for KaiNexus in 2005 while working as an emergency physician — using technology to facilitate kaizen work in the emergency department where he was practicing. The company was formally founded in 2009 and signed its first paying customer in 2011. Greg continues to practice emergency medicine two shifts per month at a local ER alongside his role at KaiNexus.
Mark Graban is from KaiNexus and is the author of "Lean Hospitals," "Healthcare Kaizen," "Measures of Success," and "The Mistakes That Make Us." He hosts the My Favorite Mistake podcast and writes at the Lean Blog.
Greg opened the session with the same kind of short reflection he gives at the start of the weekly KaiNexus team meeting. In mid-November 2020, the framing he chose was Dickensian.
The worst of times was the obvious part. U.S. COVID case numbers were completely out of control. The dynamics driving the surge were behavioral. Every person on the call had been living the same trajectory for months and didn't need it described to them again.
The best of times was the part that needed naming. The previous week had produced three pieces of news that had quietly transformed the medium-term outlook. Pfizer's interim vaccine analysis had come in at 90 percent efficacy against symptomatic disease — well above the 50 percent FDA threshold and well above the 60 to 70 percent range vaccine experts had been openly hoping for. Moderna had followed with 94.5 percent. Both used mRNA technology. Both were interim analyses based on press releases rather than published data, but the directional signal was strong, and the difference between a 70 percent vaccine and a 94 percent vaccine matters enormously for population-level outcomes.
The third piece, less covered in the news but operationally significant, was the FDA's emergency use authorization for monoclonal antibodies. Greg explained the difference for a non-medical audience. A vaccine teaches your body to produce its own antibodies, which takes weeks. A monoclonal antibody infusion gives your body antibodies directly, in a window that matters — the seven to ten days between infection and the typical point where hospitalization becomes likely. The data showed roughly a 75 percent reduction in hospitalization for high-risk patients treated early. The logistics of actually delivering it at scale were going to be a challenge, but the science had arrived.
Greg's second muse, briefer, was about trust. The standard framing — relationships require trust, business is no different — is true but incomplete. The piece that gets skipped is who goes first. His challenge to the audience: you be the first to extend trust. That's the role of leaders. Take the first step in offering trust to the people you work with, and the reciprocation tends to follow. Hold back waiting for trust to be earned before you extend it, and the standoff persists indefinitely.
The mentee polls produced a snapshot of the audience worth keeping for what it revealed about the state of CI work in late 2020.
On methodology, the room skewed toward Lean and Lean Six Sigma combinations. Daily kaizen and kaizen events were used at roughly similar rates. Strategy deployment showed up less often, which tracks with the pattern that strategy deployment tends to be a more advanced practice that organizations adopt after they've established other CI disciplines.
On the state of CI during COVID, the responses split. Some words clustered around stalled, struggling, non-existent, paused, furloughed. Other words clustered around thriving, strong, vital, fast and furious, resilient, helped us experiment. Mark noted the distribution looked potentially bimodal — organizations either pulled CI work to focus on survival, or doubled down on CI as the mechanism for managing the crisis. There wasn't much middle ground.
On whether their organization was now stronger because of CI work done during COVID, the average came in slightly above the midpoint — three out of five, skewed positive but not dramatically. Mark and Greg both noted the wording mattered. The question asked what is, not what could be. Phrased differently — would your organization be stronger with more CI — the answer would have been near unanimous.
Greg pointed to UMass Memorial Health (recently joined the KaiNexus customer base at the time of the webinar) as one of the organizations publicly reporting that the investment they had made in continuous improvement during COVID had left them performing better and better positioned for the future. That pattern — organizations that had built CI capability before the pandemic finding themselves more adaptive during it — was something the audience polling appeared to be reflecting at the individual-organization level.
The trivia segment turned into an unintentional company history. The three questions Mark had picked all happened to be milestones in KaiNexus's evolution.
First paying customer: 2011. The idea originated in 2005 when Greg was working in the emergency department. The company was formally started in 2009. The first paying customer came at the end of 2011. The 2014–2015 window was when KaiNexus raised a larger funding round and moved from "three people hacking away in a garage" to a larger team — which, in Greg's words, was "still hacking away in a garage, just with more people."
Greg is an emergency medicine physician. A meaningful fraction of audience guesses went to internal medicine, which Mark noted is the specialty most associated with the prominent figures in the Lean healthcare movement — John Toussaint of Catalysis, Gary Kaplan at Virginia Mason, others. Mark's hypothesis, which he offered as half-formed, was that internal medicine specialists are trained to think systemically about the body and that this might transfer to thinking systemically about organizations. Greg's response was characteristically less reverent: "It's a numbers game. There's a whole bunch of internal medicine doctors."
First customer was in healthcare. This held for a while — the early KaiNexus messaging and even the website was entirely healthcare-focused. The pivot to industry-agnostic came in 2013, driven by inbound demand. Mark and Greg estimated that more than half of the people reaching out to them at the time were non-healthcare prospects asking why KaiNexus only talked to hospitals. The development work to generalize the platform took about a month. The current customer mix is roughly 60 percent non-healthcare and 40 percent healthcare. The early tagline "Helping Healthcare Heal" gave way to "Making Improvement Easier," which eventually evolved into the current mission statement: spread continuous improvement.
Greg's brief state-of-KaiNexus update was honest about the year. The economic downturn had hurt. Hiring had been frozen in March out of caution. By August it was clear the company would weather the downturn without major issues, and hiring reopened. The team had added several people and was anticipating roughly 50 percent growth in the following year.
The internal challenge Greg named was less about survival and more about scaling. At ten people, everyone knew how things were done by osmosis. By twenty people, the disciplines and practices that KaiNexus helps customers build had started to be needed internally. At thirty or forty people, the lack of internal rigor was going to become a constraint.
The piece Greg emphasized about how the team had handled the uncertainty was over-communication about financial status. Many startups in 2020 weren't surviving. The KaiNexus leadership team had decided to be transparent — almost weekly — about where the company actually stood financially, and had committed publicly that nobody was going to be thrown off the proverbial boat. The over-communication mattered more than the underlying numbers. People can handle bad news. They can't handle silence.
The first Ask Us Anything question came from Jason: his team was trying to start a CI culture and having trouble getting staff buy-in.
Mark's opening framing was diagnostic. What looks like resistance is usually something more specific. Historical remnants of a previous suggestion-box system that didn't work. Fear that speaking up with ideas brings negative attention rather than positive. Communication patterns that signal participation isn't actually welcome despite what's being said publicly. The fix isn't to push harder on people to participate. The fix is to figure out what they've already learned about whether participation is safe and worthwhile in this specific organization.
Greg's answer was the one he said he would have given differently fifteen years ago. Early in his journey, his answer would have been "just start." Now, fifteen years in, he sees the role of leaders as the actual determinant. The cultures that succeed, sustain, and spread are driven by leaders at the organization — not necessarily only the C-suite, but local leaders too, at department and team level.
The leadership behaviors that matter aren't only the cheerleading. They're the translation of cheerleading into actually doing improvement work. A leader participating in a gemba walk. A leader doing their own standard work. A leader asking team members, in an inquisitive rather than accusatory way, "what improvement have you participated in recently?" Done as a discipline rather than a one-time performance, these behaviors signal to the rest of the organization that improvement is real rather than rhetorical.
If you're not in a leadership role and trying to drive a CI culture from below, Greg's advice was practical. Put your energy into engaging a leader. Get one of them bought in. Working with a leader who is genuinely participating is structurally easier than trying to build CI culture around leaders who treat it as someone else's job.
Mark added a complementary point about why apparent buy-in problems sometimes aren't buy-in problems at all. People who are stuck on coming up with big, million-dollar ideas often produce no ideas because they've put too much pressure on themselves. The pressure stifles the creativity rather than focusing it. Making a game of it — asking what's the smallest improvement someone could come up with — can unstick the process. Once the wheels are turning, the larger improvements come naturally.
Greg pulled in James Clear's "Atomic Habits" as the supporting framework. The question of starting a CI culture is, structurally, the question of building an organizational habit of thinking about improvement in a disciplined and repetitive way. Habit science is clear that the way to build a habit is not to attempt everything at once — bottom-up improvement, top-down strategy, daily kaizen, kaizen events, strategy deployment — all simultaneously. The way to build a habit is to start small with something achievable, make it routine, and layer on additional practices once the first one is sticking.
For bottom-up improvement specifically, the question Greg recommended asking people is some variation of "what bugs you" or "what would make your life easier if we tackled it" — followed by coaching them to the low-cost, low-risk, ten-feet-around-you range. The big systemic problems will come later. The first wins need to be close enough to the work that people can see them happen.
Mark added a book recommendation that's worth pulling out: Robert Maurer's "The Spirit of Kaizen," which Mark mentioned is sold by Toyota at their visitor center in Japan — a quiet endorsement that says something. Maurer is a psychologist, and his books address the psychology of getting started rather than the methodology of doing improvement work.
The follow-up question from Alex was the one most internal CI teams want to ask of any CI software company. How do you actually use this internally?
Greg's answer was honest about both what KaiNexus does well and where the team has had to build its own discipline over time.
For bottom-up improvement, the team has a system. Improvements are submitted into KaiNexus with a small number of fields. The leadership team meets weekly to make sure things aren't slipping through the cracks. This pattern has been in place for years and works reliably at the team's scale.
For project work — strategic projects, in Greg's terminology, meaning work that's bigger than a small improvement and not part of someone's everyday job — the KaiNexus team has had to develop more rigor as it has grown. Eight or ten years ago the platform was focused on bottom-up improvement and project work happened informally. As the team grew from eight or ten people to twenty-five or thirty, the informal approach started to break down. The team started using KaiNexus for project tracking around 2016 or 2017, then expanded the rigor in the past couple of years.
The lesson Greg pulled out from 2020 specifically: strategic projects do better with at least two people assigned. Projects with a single owner tend to stall until someone — often Greg himself — intervenes with a "what's happening with this?" conversation. Projects with multiple people involved produce more energy and more reliable progress. The team had recently moved a team member into a role helping organize the strategic project portfolio — not doing all the projects, but making sure the team knows which projects are underway, that work isn't being duplicated, and that the team isn't trying to do too many things at once.
Data management at KaiNexus runs through the platform as well. Every division (marketing, development, customer experience, product, sales) has its own board reporting on data. The team tries to keep each board to no more than six or seven metrics — recognition that more data on a board doesn't produce more attention to the data. KaiNexus also uses the platform in less standard ways: managing customer relationships, tracking bug reports, and other internal applications that take advantage of the workflow and visibility capabilities.
The implicit lesson across the eating-your-own-dog-food discussion: continuous improvement infrastructure works the same way for the company that builds it as for the customers it serves. The disciplines required are the disciplines required. The growth pains a thirty-person CI software company experiences are the growth pains any thirty-person team experiences. There's no exemption for being the experts on the topic.
A question from Felipe surfaced one of the more nuanced threads of the conversation. What's the best way to recognize team members' contributions to improvement?
Greg's answer started simple. Literally recognize them. The KaiNexus team has built a habit of giving shout-outs in weekly meetings, and the effect is real — both for the people being recognized and, in Greg's experience, for the leaders who occasionally receive a shout-out themselves and remember how much it matters.
The piece Greg emphasized was visibility. There's a difference between pulling someone aside to tell them you appreciated their work and saying it in front of the team. Both have value. The public version has more.
Mark added a caveat that's worth keeping. Most people do prefer being recognized publicly. Some people genuinely hate it. The default should be public recognition with awareness that for at least some people, that default produces the opposite of the intended effect.
Then Mark moved to the broader frame. There's often too much emphasis in the conversation on rewards relative to recognition. Tapping into people's intrinsic motivation — making their work less frustrating, more interesting, more fulfilling, more directly connected to the customer they're serving — produces results that no amount of cash incentive or quota-based system can replicate. Recognition costs nothing other than the time to do it. A good culture doesn't have to force improvement through targets, quotas, or minimums. Forcing it tends to make it dysfunctional.
Greg recommended Daniel Pink's "Drive" for anyone wanting the science behind the claim that intrinsic motivation outperforms extrinsic motivation. Pink's book surveys the social science research showing why traditional carrot-and-stick approaches produce worse results than approaches that tap into autonomy, mastery, and purpose. Before installing a recognition program or behavioral incentive system, read Pink.
The follow-up question from Chris was the one that takes this conversation seriously: how do you balance the need for recognition without making it contrived, given that there may be diminishing returns?
Mark's answer was that this is genuinely the art of leadership. Watch the feedback loops from your team. Notice how they're responding. Notice when the response is changing. Lean in and try to understand what's causing the shift. Adjust your approach. One size doesn't fit all, and the one size that fits one team won't necessarily fit the same team six months later. The one principle that does apply universally is authenticity. People can detect inauthentic recognition immediately, and inauthentic recognition is worse than no recognition at all.
Chris asked what process improvement Mark and Greg thought organizations should have been doing all along but the pandemic actually pushed into the mainstream.
Mark's example was telemedicine. He had talked to multiple health systems whose three-year or five-year roadmaps for video visits had been implemented in two or three weeks during the early pandemic. The capability had been there. The motivation to actually deploy it hadn't. When the motivation arrived, the timeline compressed dramatically. The question worth sitting with is which other three-year roadmaps in your organization could be a two-week implementation if the motivation showed up.
Greg's broader framing was about the relationship between perfect and good. The pandemic forced organizations to embrace good enough — to ship things that worked rather than waiting for the polished version, and to revisit them later when conditions allowed. The realization that the things they had shipped under crisis conditions were producing real value led naturally to the question of why this approach hadn't been the default all along. The answer, of course, is that perfect-as-the-enemy-of-good is a habit that organizations build deliberately over years. The pandemic broke the habit by force. Whether the new habit persists is a question that will play out over the next decade.
Mark added a small example along the lines of personal process improvement. In parts of Asia, particularly Japan, wearing a mask in public has been common for years — partly as a courtesy to avoid spreading germs to others. He wondered whether the practice of wearing a mask on an airplane might persist post-pandemic for the same reason. The cost is minimal. The benefit, if it reduces the frequency of getting sick after flights, is real. Some people will be happy to get rid of masks permanently. Others might keep wearing them on planes specifically because the cost-benefit math works out, and because the social acceptance of mask-wearing has shifted permanently.
Kathy asked about mapping process flows when you don't have physical whiteboards, sticky notes, or flip charts — and noted that virtual whiteboard tools, in her experience, were hard to use for quick mapping.
Mark's response was practical. Google Docs and Google Sheets work better than they should for collaborative real-time work, because multiple people can edit simultaneously and the friction is low. He had done some collaborative mapping during the early pandemic using Google Slides. Even a plain document with bulleted process steps can serve as a first pass — get the structure on the page, then someone can come back later and draw the flow if needed.
Greg's response was that the KaiNexus team was doing serious deep thinking on this topic — and left it at that.
The conversation across this session keeps returning to one operational point: continuous improvement at scale needs infrastructure, and the infrastructure has to be the kind that supports the actual behaviors of CI work rather than getting in the way of them.
The bottom-up improvement system Greg described — submissions into a structured format, a weekly meeting to make sure nothing falls through the cracks, visibility to leadership — is the kind of thing that works on a small team using almost any tool. It also stops working past a certain scale unless the structure is deliberate. The KaiNexus team's own experience of moving from ten people to thirty is a small-scale version of what customer organizations experience moving from a single CI pilot to enterprise-scale improvement work. The same pattern repeats. The informal approach that worked early breaks down. The disciplines and practices that the platform was built to support become necessary internally as the team grows past the point where institutional knowledge alone can carry the operational load.
The recognition discussion has a structural component as well. Recognition that depends on a single leader remembering to give shout-outs in every weekly meeting is fragile. Recognition that's woven into the visibility of the work — where impacts get tracked, where ideas get acknowledged in the flow rather than as separate events, where leadership can see what's happening across the organization and respond to it — sustains across leadership changes and team growth in a way that purely personal practices can't.
The eating-your-own-dog-food discussion was the most direct statement of what the platform does and what it doesn't do. KaiNexus the company uses KaiNexus the platform for bottom-up improvement, strategic project tracking, data dashboards by division, customer relationship management, bug tracking, and a variety of other workflows. The platform doesn't make the improvement work happen. The disciplines of weekly leadership meetings, multi-person project ownership, board visibility, and the actual leadership behaviors that drive CI culture are what produce the results. The platform is the infrastructure that keeps those disciplines visible and trackable as the organization grows.
None of this changes what Mark and Greg were teaching across the session. The leadership work is the leadership work. The trust extension is the trust extension. The recognition is the recognition. The discipline of starting small and building up is the discipline of starting small and building up. What infrastructure does is keep the work from getting lost as the organization scales, and make it possible for leaders to see and respond to what's happening across teams without having to be physically present at every conversation.
How do you start a continuous improvement culture when staff buy-in is low? Diagnose the resistance before pushing harder. What looks like buy-in problems is usually something specific — a previous suggestion-box system that didn't work, fear that speaking up brings negative attention, communication patterns that signal participation isn't actually welcome. Engage a leader who will not only champion the work verbally but also do their own improvement work visibly. If you're not in a leadership role, put your energy into getting a leader bought in before trying to drive the culture from below. Start small. Don't ask people for million-dollar ideas. Ask what bugs them about their work and coach them to low-cost, low-risk, ten-feet-around-you improvements.
Why does leader behavior matter so much for CI culture? Because organizations watch what leaders do, not what they say. A leader who champions CI in town halls and emails but never participates in a gemba walk, never asks team members what improvements they've worked on, and never does their own improvement work teaches the organization that CI is rhetoric. A leader who consistently translates the rhetoric into participation teaches the organization that CI is real. The behaviors compound over time. They're also why "just start" advice without leader engagement tends to produce short-lived enthusiasm rather than sustained culture change.
What does Greg mean by leaders going first on trust? That trust in any relationship — including business relationships — requires someone to take the first step. The default for many leaders is to wait for trust to be earned before extending it. The result is a standoff that can last indefinitely. Greg's challenge to leaders is to be the first to extend trust to the people they work with. The reciprocation tends to follow. The standoff approach almost never produces what the standoff is theoretically protecting against.
Should you start CI with bottom-up improvement or top-down strategy deployment? Start with one. Don't start with everything. Habit science (Greg referenced James Clear's "Atomic Habits") suggests that the way to build a habit is to make it small enough to actually do consistently, then layer on additional practices once the first is sticking. Trying to deploy bottom-up improvement, top-down strategy, daily kaizen, kaizen events, and strategy deployment all at once tends to produce none of them. For most organizations, starting with bottom-up improvement — asking people what bugs them, coaching to low-cost low-risk fixes — produces early wins that build the cultural foundation for more sophisticated practices later.
How does the KaiNexus team use KaiNexus internally? For bottom-up improvements (submitted as "improvements," reviewed in a weekly leadership meeting), for strategic project tracking (more rigor added as the team grew past 20 people), for division-level data dashboards (capped at six or seven metrics per board), for customer relationship management, for bug tracking, and for various other workflows that take advantage of the platform's visibility and accountability features. The team has had to build its own internal discipline as it has grown — the same pattern customers experience as their CI work scales.
What did Greg learn in 2020 about strategic projects? That projects with a single owner tend to stall, while projects with at least two people assigned generate more energy and reliable progress. The KaiNexus team had moved a team member into a role helping organize the strategic project portfolio — not doing all the projects, but making sure the team knew which projects were underway, that work wasn't being duplicated, and that the team wasn't trying to do too many things at once.
What's the most important thing about recognition? Authenticity. Most people prefer public recognition over private — but some people genuinely hate public recognition, and the default needs awareness of that. Beyond the individual variation, the bigger point is that recognition is structurally different from rewards. Recognition costs nothing other than the time to do it. Rewards (cash incentives, quotas, targets) often produce dysfunctional results because they crowd out intrinsic motivation. Daniel Pink's "Drive" surveys the science on why intrinsic motivation outperforms extrinsic motivation across most knowledge-work contexts.
How do you avoid recognition becoming contrived? Watch the feedback loops from your team. Notice how they're responding. Notice when responses change. Adjust your approach when you see the response shifting. One size doesn't fit all, and one size that fits a team today may not fit the same team in six months. The one principle that applies universally is authenticity — people detect inauthentic recognition immediately, and inauthentic recognition is worse than no recognition at all.
What process improvements did the pandemic accelerate? Telemedicine is the clearest example in healthcare. Roadmaps that had been three or five years long were implemented in two or three weeks once the motivation arrived. The broader pattern is that organizations were forced to embrace "good enough" rather than waiting for perfect — and many of them discovered that the things they shipped under crisis conditions were producing real value. Whether the new habit of shipping good-enough versions and iterating persists is a question that will play out over time, but the precedent has been set.
How do you map processes virtually when whiteboard tools are clumsy? Mark recommended Google Docs, Google Sheets, and Google Slides for collaborative real-time work — the friction is low and multiple people can edit simultaneously. Even a plain document with bulleted process steps can serve as a first pass before drawing any flow diagrams. The bigger principle is to optimize for actually getting the mapping done rather than for the polish of the visual output. The map exists to support the conversation. If the tool gets in the way of the conversation, simplify the tool.
What's the difference between a vaccine and a monoclonal antibody? A vaccine introduces something into the body that triggers the person's own immune system to produce antibodies — useful for prevention before exposure. A monoclonal antibody is an infusion of antibodies directly into the bloodstream, which bypasses the two-to-three-week window the body normally needs to produce its own antibodies. Monoclonal antibodies are typically used for people who are already infected and at high risk, early in the course of illness, to prevent progression to hospitalization. The November 2020 EUA showed roughly a 75 percent reduction in hospitalization for high-risk patients treated early.
What is "Letters to Humans"? Greg's pandemic-era email newsletter, started in March 2020 as a daily letter to family and friends explaining what was about to happen with COVID. It eventually moved to a weekly Thursday cadence and covers the science of the pandemic without political commentary. Greg has no commercial agenda for the newsletter — it exists to communicate meaningful scientific information to people who want to understand what's happening. Available at letterstohumans.com.
![[Watch Now] The Why, How, and What of Continuous Improvement](https://no-cache.hubspot.com/cta/default/326641/060c8cc9-44ad-4fd1-ad7b-7b4ab71c7632.png)
.svg)
Copyright © 2026
Privacy Policy