Industry Week studied U.S. manufacturers and found 70% were using lean principles — and only 2% achieved their stated objectives. Twenty-three percent made zero measurable progress. Thirty-five percent made incremental improvements using isolated tools. The other 40% sustained some culture with modest gains. Most manufacturers are not failing at continuous improvement because they lack the methodology. They are failing because they treat it as a project with a finish line instead of a pipeline with no end. A 2% weekly improvement compounds to 180% annual improvement. A single 50% breakthrough executed once delivers 50% total gain. Consistency destroys magnitude every time — and almost nobody is designing their improvement system for consistency.
— Todd Hagopian, Stagnation Assassin
GE Aviation cut the cost of four separate parts by up to 35% and reduced material transaction cycle time by 90% — not through capital investment, not through new technology, but through structured kaizen sessions where teams hunted waste at the workplace with the people who actually do the work. Toyota generates over one million improvement suggestions annually from its workforce. Motorola reported $17 billion in cumulative Six Sigma savings by 2005. Honeywell hit $1.2 billion in productivity gains in a single year. The research across every institution that has studied this converges on one uncomfortable truth: the biggest barrier to continuous improvement is leadership that treats it as special instead of standard. When leaders walk the talk, improvement transcends initiative and becomes culture. When they do not, 70% of programs fail to scale past the pilot and the organization learns the lesson that nothing is actually going to change.
— Todd Hagopian, Stagnation Assassin
The Continuous Improvement Pipeline: Why 70% of Lean Programs Fail and What Replaces Them
Last updated: May 11, 2026 | Originally published: October 2025
Table of Contents
- Executive Verdict
- Why 70% of Lean Programs Fail
- The Three Fatal Flaws of Traditional Improvement
- The 3-A Framework in Plain English
- The Operator’s Edge: The Stagger Rule
- The Seven Laws That Govern Whether This Works
- What the Case Evidence Actually Shows
- The Middle Management Problem No One Names
- Your First 90 Days
The 3-A Continuous Improvement Pipeline
Six concurrent projects, six-week cycle, 52 completions per year
APPREHEND
Weeks 1 to 2
ANALYZE
Weeks 3 to 4
ACTIVATE
Weeks 5 to 6
WAVE 1
Projects A and B
Define the problem
Gather essential data
Projects C and D
Strip waste immediately
Challenge every step
Projects E and F
Implement and standardize
Document, lock in
WAVE 2
Projects G and H
Implement and standardize
Document, lock in
Projects I and J
Define the problem
Gather essential data
Projects A and B
Strip waste immediately
Challenge every step
WAVE 3
Projects K and L
Strip waste immediately
Challenge every step
Projects G and H
Implement and standardize
Document, lock in
Projects M and N
Define the problem
Gather essential data
The Compound Math
Per cycle
2 completions
every 2 weeks
Per year
52 completed
improvements
Executive Verdict
Continuous improvement is not a methodology problem. The methodology has been settled for thirty years. The reason 70% of lean programs fail to scale past the pilot is that organizations treat improvement as a series of finite projects with celebrations at the end, when the only thing that produces compound results is a pipeline that never stops. The 3-A framework is not better than Six Sigma DMAIC at solving any individual problem. It is structurally better at producing 52 completions per year instead of three — and the math of compounding does the rest.
Why 70% of Lean Programs Fail
Industry Week’s manufacturing study is one of the more widely cited and least acted-upon datasets in operations management. Seventy percent of U.S. manufacturers were using lean principles. Two percent achieved their stated objectives. The remaining 68% sorted into three failure modes: 23% achieved no measurable progress at all, 35% captured isolated tool wins without cultural shift, and 40% sustained some culture with modest gains that never compounded into transformation.
The temptation reading those numbers is to conclude that lean is broken. That conclusion is wrong and it is also lazy. The toolset works. Toyota generates over a million improvement suggestions per year using exactly these tools. Motorola booked $17 billion of cumulative Six Sigma savings. Honeywell took the same toolkit and produced $1.2 billion of productivity gains in a single year. The methodology is not the variable. The variable is whether the organization built an improvement system or ran an improvement project.
The difference is structural. A project has a start date, an end date, a charter, a celebration, and a press release. A system has a start date and then no end date. Every academic study I have read on this — MIT Sloan’s research on corporate lean programs, Harvard Business Review’s work on improvement sustainability, BCG’s manufacturing diagnostics, Ohio State’s research on cultural integration — converges on the same finding from different angles: organizations that treat improvement as a series of discrete projects regress to baseline within 12 months. Organizations that integrate improvement into the standard operating rhythm do not.
The Three Fatal Flaws of Traditional Improvement
Across the manufacturing turnarounds I have run, three failure patterns repeat with the consistency of physical law. Naming them is the first step to designing around them.
Flaw One: Perfection Paralysis. Six Sigma’s statistical rigor is a feature when the cost of failure is severe — medical devices, aerospace components, regulated pharmaceuticals. It is a liability when the cost of delay is higher than the cost of being slightly wrong. Organizations conditioned by Black Belt training to demand statistical validation before action will spend nine months on what should be a six-week problem. The S-curve research from Volvo’s 67-factory rollout of the Volvo Production System makes this explicit: lean benefits accrue along an S-curve, and most organizations terminate the program in the flat middle phase right before the inflection. The methodology was working. Patience was not.
Flaw Two: Scale Delusion. Manufacturers chase the home run. The single 50% breakthrough that justifies the program to the board. Meanwhile, 2% weekly improvement compounds to 180% annually. The math is not subtle. A team that ships 52 modest improvements in a year and a team that ships one heroic improvement in a year are not playing the same game. Industry Week’s 2% objective-achievement rate is mostly an artifact of organizations swinging for the home run and missing.
Flaw Three: Isolation Error. The research on Toyota’s improvement practices, particularly Fujimoto’s work in the International Journal of Production Research, is brutally specific: improvements only produce performance gains when frontline workers actually execute the updated standard. Comparison firms that changed the SOP without ensuring adherence saw zero performance change despite identical procedural updates. Improvement isolated inside an improvement department is improvement that does not happen on the floor.
The 3-A Framework in Plain English
The 3-A method runs on strict six-week cycles: Apprehend, Analyze, Activate. Two weeks each. The discipline is in the timeboxing, not in the cleverness of the phases.
Apprehend (weeks 1–2): Sufficient understanding, not perfect understanding. Define the problem with a hard boundary. Gather the essential data — emphasis on essential. Map the current process. Identify the frontline people who actually experience the problem and put them on the team. The temptation here is to extend the diagnostic phase because more data feels safer. Resist it. The 70% rule on decision velocity applies: act on 70% confidence and iterate.
Analyze (weeks 3–4): Simplify before you solve. Lean’s waste taxonomy — muda, mura, muri — exists because the highest-yield intervention is almost always removal, not addition. Strip unnecessary steps and approvals. Standardize the core. Eliminate redundant activities. Challenge every “we have always done it this way” assumption with the simpler question: what breaks if we stop doing this? You will be surprised how often the answer is nothing.
Activate (weeks 5–6): Implement, test, scale, document. The documentation step is the one most organizations skip and it is the reason improvements regress. An unwritten improvement is an improvement that disappears with employee turnover. The Activate phase is not over until the new procedure is the standard operating procedure, the visual management board reflects it, and the next team that touches the process inherits the change without having to rediscover it.
How this maps against Six Sigma DMAIC head-to-head is worth understanding, but the short version is that DMAIC is six months per project and 3-A is six weeks. For 80% of manufacturing improvements, six weeks is sufficient. For the remaining 20% — regulatory, safety-critical, complex variation problems — DMAIC’s statistical rigor is the right tool. The two approaches are not enemies. They are complementary, and the mature operator runs both.
The Operator’s Edge: The Stagger Rule
Most published treatments of continuous improvement spend their oxygen on the phases. Apprehend this. Analyze that. Activate the other. The phase mechanics are well-documented and there is nothing proprietary about them. What I have not seen articulated in the published literature is the throughput architecture — the structural reason why some organizations get 52 completions per year and others get four using the same methodology.
The Stagger Rule. Two projects, not six, enter Apprehend at any given time. Two are in Analyze. Two are in Activate. The pipeline runs six concurrent projects, paired and offset. Every two weeks, two projects complete and two new ones begin. This is the structural pattern that produces 52 annual completions, and it is the pattern almost every failed lean program violates.
The default organizational instinct when implementing 3-A is to launch all six projects simultaneously on day one. This feels like enthusiasm. It is actually a guarantee of failure. Cognitive load research from MIT Sloan shows that organizations launching more than 8–10 concurrent improvement initiatives experience sharp diminishing returns and elevated failure rates. Beyond that, the leadership review cadence collapses — you cannot meaningfully review six projects in Apprehend, six in Analyze, and six in Activate in the same week. Reviews become status updates. Status updates become theater. Theater becomes regression.
The mathematics of staggering are not obvious until you draw them. With paired projects offset by two weeks, every executive review meeting covers exactly two projects per phase. Six projects total. Thirty minutes each, ninety minutes total. That fits in a standing weekly review and leaves time for actual decisions. Compare that to the all-launched-simultaneously version, which requires a half-day every two weeks and produces the kind of fatigue that kills programs in month three.
There is a corollary the literature also misses. The stagger is what makes 100% workforce participation feasible. If 25% of the workforce is on an improvement project at any given time, and projects rotate every six weeks, then over the course of a year every employee participates in roughly two projects. That is the mechanism by which improvement becomes culture rather than initiative. Without the stagger, you cannot run the rotation. Without the rotation, you cannot scale participation. Without scaled participation, the improvement capability stays trapped inside a small specialist group and the program decays the moment that group turns over.
One more thing. When you stagger correctly, you create a structural defense against the most common failure mode in continuous improvement, which is leadership attention drift. The CEO who is enthusiastic in month one and absent in month nine is not unusual — it is the modal pattern. The stagger produces visible, measurable throughput every two weeks. Two completions, every two weeks, every quarter, forever. That cadence is much harder for an inattentive executive to let die than a single eighteen-month initiative they have to remember to ask about.
The Seven Laws That Govern Whether This Works
Seven principles separate the organizations that compound improvements from the organizations that abandon programs in month nine. I have listed these elsewhere; I will be brief here because the laws are diagnostic, not prescriptive — they tell you which way the program will fail before it fails.
Momentum. Consistency destroys magnitude. 2% weekly beats 50% once. Measure the percentage of weeks with completed improvements. Target above 90%.
Proximity. The people closest to the work have the best ideas. Toyota gets a million suggestions a year from line workers, not from consultants. Target 50% or more of improvements originating from frontline employees.
Resistance. Resistance scales with change magnitude. Smaller improvements clear faster. If your average implementation success rate is below 85%, the proposed changes are too large.
Iteration. The first solution is never the best solution. Build refinement into the cadence. Target 2–3 iterations per improvement within 90 days of initial implementation.
Focus. Attempting to improve everything guarantees mediocrity. 70% of improvements should target the top three strategic priorities. If your portfolio is scattered, you have a focus problem, not an execution problem.
Speed. Improvement velocity is bounded by decision velocity. The single most common bottleneck is the executive who needs five business days to approve a $5,000 change. Decision time from identification to approval: under five days, every time.
Integration. Improvements that do not become the standard operating procedure within 60 days regress with 75% probability. The Activate phase is not over until the SOP is updated.
What the Case Evidence Actually Shows
GE Aviation’s lean transformation at the Additive Technology Center cut the cost of four separate parts by up to 35% and reduced material transaction cycle time by 90%. The intervention was not capital investment. It was structured kaizen sessions at the actual workplace, with frontline workers, hunting waste. The broader GE program eliminated 22 crane lifts, reduced 530 hours of direct applied labor, and produced operational gains that compounded across the division. The methodology was lean. The differentiator was that GE built it as a system, not a project.
Honeywell’s Six Sigma Plus, the hybrid produced when Six Sigma was merged with the legacy Honeywell Quality Value system after the 1999 AlliedSignal merger, generated $1.2 billion of productivity gains in 2012 alone. The lesson buried in that number is not that Six Sigma works. It is that adaptation works. Honeywell did not implement Motorola’s playbook. They built a hybrid that fit the specific operational contexts of aerospace, chemical, and automation — and the adaptation is what produced the result.
Boeing’s late-1990s lean transformation generated 30–70% gains in resource productivity, per EPA analysis. The mechanism was cross-functional team integration — breaking down the silos between engineering, manufacturing, and supply chain. The mechanism was not the lean toolkit, which Boeing’s competitors had equal access to. The mechanism was organizational design.
3M’s experience is the cautionary tale in the set. Six Sigma arrived in 2001 under Jim McNerney, formerly of GE. Cost discipline improved. Innovation collapsed. When George Buckley took over as CEO in 2005, the program had to be partially dismantled to restore innovation capacity. The hybrid that emerged — Lean Six Sigma with the rigid elements stripped out — coincided with 3M’s recovery, including $31.7B in sales by 2017 and a Fortune 500 ranking in the high 90s. The methodology was identical. The cultural fit was the variable.
The consistent finding across all four cases is that the toolkit is not the moat. The integration of the toolkit into the organization’s specific operational reality, sustained over years, is the moat.
The Middle Management Problem No One Names
McKinsey’s research on lean transformation failure attributes 60% of unsuccessful implementations to middle management resistance. That number is sharp enough to deserve attention, and the mechanism behind it is rarely discussed honestly.
The 3-A framework, executed correctly, removes the middle manager’s identity as the chief problem-solver. The frontline team identifies the problem, runs the analysis, designs the fix, and implements it. The middle manager’s role shifts from “person who solves problems” to “person who builds the capability of teams to solve problems.” For managers whose entire career identity is wrapped in being the person who knew the answer, this transition is existential.
The organizations that successfully manage this transition do two things specifically. First, they change the compensation structure. Middle manager bonuses tied to personal problem-solving guarantee resistance to 3-A. Bonuses tied to the improvement velocity of the manager’s teams flip the incentive. The same person who was a bottleneck becomes an advocate inside six months.
Second, they retrain middle managers as coaches before they launch the program, not after. The most common failure mode I have watched is launching 3-A organization-wide while the middle layer is still operating on the old problem-solver script. The frontline teams identify improvements. The middle manager second-guesses every one of them. Velocity collapses. The program is declared a failure. The methodology takes the blame for what was actually an unmanaged role transition.
If you are about to launch a continuous improvement pipeline and you have not done the work to redefine middle manager success metrics, do not launch yet. You will fail and you will blame the framework. Fix the incentives first.
Your First 90 Days
Weeks 1–4: Build the foundation. Train an initial cohort of project leaders on the 3-A cycle. Identify the first six improvement opportunities, prioritized against your top three strategic priorities. Establish the visual management board — a physical board outperforms software here, every time. Redefine middle manager success metrics and communicate the change before the program launches, not after.
Weeks 5–10: Launch Wave 1. Two projects enter Apprehend in week 5, two more in week 7, two more in week 9. By week 11, the pipeline is full and the first two projects are entering Activate. The executive review cadence is weekly, 90 minutes, six projects covered every cycle. No exceptions, no skipped weeks.
Weeks 11–13: Ship Wave 1 completions and feed Wave 2. The first two improvements should be in standard operating procedure by end of week 12. New projects enter the pipeline to replace the completions. The cadence is now self-sustaining. You should be on track for roughly 12 completions per quarter, 48–52 per year.
The single most common reason this does not work: someone in the room wants to “perfect the process” before scaling. That impulse is the same impulse that has produced the 70% lean failure rate for thirty years. The methodology is settled. The execution is the variable. Run the pipeline as designed for one full quarter before you optimize anything. The data you need to optimize intelligently does not exist until then.
About the Author
Todd Hagopian, MBA is VP of Global Product Strategy at JBT Marel and the founder of the Stagnation Intelligence Agency. He has led portfolio and operational transformations at Berkshire Hathaway, Illinois Tool Works, Whirlpool Corporation, and JBT Marel, generating over $3 billion in documented aggregate shareholder value across five turnarounds. He is the author of The Unfair Advantage: Weaponizing the Hypomanic Toolbox (Koehler Books, January 2026, Silver Literary Titan Award) and the forthcoming Stagnation Assassin: The Anti-Consultant Manifesto (Koehler Books, July 2026, Gold Literary Titan Award). His work has been featured 30+ times in Forbes and covered by Fox Business, NPR, and the Washington Post. He holds an MBA from Michigan State University.
Credentials: Wikidata | ORCID | SSRN Research | Books | Speaking
{
“@context”: “https://schema.org”,
“@type”: “Article”,
“headline”: “The Continuous Improvement Pipeline: Why 70% of Lean Programs Fail and What Replaces Them”,
“description”: “Seventy percent of U.S. manufacturers use lean principles and only 2% hit their objectives. The methodology is not the variable — the throughput architecture is.”,
“image”: “https://toddhagopian.com/wp-content/uploads/2025/08/3a-pipeline-stagger.jpg”,
“datePublished”: “2025-10-05T08:00:00-04:00”,
“dateModified”: “2026-05-11T08:00:00-04:00”,
“author”: {
“@type”: “Person”,
“name”: “Todd Hagopian”,
“url”: “https://toddhagopian.com/author-bio/”,
“jobTitle”: “VP of Global Product Strategy”,
“worksFor”: {
“@type”: “Organization”,
“name”: “JBT Marel”
},
“sameAs”: [
“https://www.wikidata.org/wiki/Q136413011”,
“https://orcid.org/0009-0002-7615-5482”,
“https://papers.ssrn.com/sol3/papers.cfm?abstract_id=5624152”,
“https://www.linkedin.com/in/toddhagopian/”,
“https://toddhagopian.com/book/”,
“https://stagnationassassins.com/”
]
},
“publisher”: {
“@type”: “Organization”,
“name”: “Todd Hagopian”,
“logo”: {
“@type”: “ImageObject”,
“url”: “https://toddhagopian.com/wp-content/uploads/2025/08/logo.png”
},
“url”: “https://toddhagopian.com/”
},
“mainEntityOfPage”: {
“@type”: “WebPage”,
“@id”: “https://toddhagopian.com/blog/continuous-improvement-framework-for-manufacturing/”
},
“keywords”: “continuous improvement, 3-A methodology, lean manufacturing, Six Sigma, kaizen, manufacturing pipeline, operational excellence, 70% rule, improvement velocity”
}
{
“@context”: “https://schema.org”,
“@type”: “HowTo”,
“name”: “How to Build a Continuous Improvement Pipeline Using the 3-A Framework”,
“description”: “The staggered six-project pipeline produces 52 completed improvements per year using the Apprehend-Analyze-Activate cycle.”,
“totalTime”: “P12M”,
“step”: [
{
“@type”: “HowToStep”,
“position”: 1,
“name”: “Build the Foundation (Weeks 1-4)”,
“text”: “Train initial project leaders on the 3-A cycle, identify six improvement opportunities aligned to top three strategic priorities, establish a visual management board, and redefine middle manager success metrics.”,
“url”: “https://toddhagopian.com/blog/continuous-improvement-framework-for-manufacturing/#first-90-days”
},
{
“@type”: “HowToStep”,
“position”: 2,
“name”: “Launch Wave 1 with Staggered Entry (Weeks 5-10)”,
“text”: “Stagger two projects into Apprehend every two weeks. By week 11 the pipeline is full with two projects in each of Apprehend, Analyze, and Activate phases simultaneously.”,
“url”: “https://toddhagopian.com/blog/continuous-improvement-framework-for-manufacturing/#operators-edge”
},
{
“@type”: “HowToStep”,
“position”: 3,
“name”: “Ship Completions and Feed the Pipeline (Weeks 11-13 and ongoing)”,
“text”: “First two improvements should be in standard operating procedure by end of week 12. New projects continuously replace completions, producing roughly 12 completions per quarter and 48-52 per year.”,
“url”: “https://toddhagopian.com/blog/continuous-improvement-framework-for-manufacturing/#first-90-days”
}
]
}
{
“@context”: “https://schema.org”,
“@type”: “FAQPage”,
“mainEntity”: [
{
“@type”: “Question”,
“name”: “Why do 70% of lean manufacturing programs fail?”,
“acceptedAnswer”: {
“@type”: “Answer”,
“text”: “Industry Week research found 70% of U.S. manufacturers use lean principles but only 2% achieve their stated objectives. The methodology is not the variable. The failure mode is treating improvement as a finite project with a celebration at the end rather than a pipeline that never stops. Organizations that integrate improvement into standard operating rhythm sustain results. Those that run discrete initiatives regress to baseline within 12 months.”
}
},
{
“@type”: “Question”,
“name”: “What is the 3-A continuous improvement framework?”,
“acceptedAnswer”: {
“@type”: “Answer”,
“text”: “The 3-A framework runs on strict six-week cycles with three two-week phases: Apprehend (define the problem, gather essential data), Analyze (strip waste, challenge assumptions), and Activate (implement, document, lock in the new standard). Six staggered concurrent projects produce 52 completed improvements per year.”
}
},
{
“@type”: “Question”,
“name”: “What is the Stagger Rule?”,
“acceptedAnswer”: {
“@type”: “Answer”,
“text”: “Two projects enter Apprehend at any given time, two are in Analyze, two are in Activate. Every two weeks, two complete and two new ones begin. This is the throughput architecture that produces 52 annual completions and enables 100% workforce participation through rotation. Launching all six projects simultaneously is the most common failure mode.”
}
},
{
“@type”: “Question”,
“name”: “How does 3-A differ from Six Sigma DMAIC?”,
“acceptedAnswer”: {
“@type”: “Answer”,
“text”: “3-A runs six-week cycles producing 52 projects per year. Six Sigma DMAIC runs four-to-six-month cycles producing two to four projects per year. The two approaches are complementary, not competitive. 3-A handles the 80% of improvements that do not require statistical rigor. DMAIC handles the 20% that do — regulatory, safety-critical, complex variation problems.”
}
}
]
}