Home > The Lean Post> Rethinking the Model Line

Rethinking the Model Line

by James P. Womack
April 19, 2018

Rethinking the Model Line

by James P. Womack
April 19, 2018 | Comments (11)

It was love at first sight. The first time I met a model line I fell in love with it. In 1993 I was on a tour at a grimy car-parts factory in the US Midwest with Toyota TSSC advisors, who exposed me to the concept of the model line, which was pioneered in the late 1960s by Taiichi Ohno and the team he created to spread the Toyota production philosophy throughout the Toyota supplier group in Japan.

At this factory the TSSC team had directed the movement of machines out of process villages to create a cell composed of technologies (molding, paint, final assembly, and test) that did not normally live together. In so doing they had created single-piece-flow with dramatically reduced lead times, reduced cost, and higher quality. This was for one product family amidst the chaos of a mass production facility with many product families.

And yet. In walking around the rest of the factory I noted that nothing else had changed. Work elsewhere proceeded in the traditional way, and did so in a deep dim gloom compared with the one new cell that had introduced bright lighting. The performance improvements in the model line were so compelling that I had mistakenly assumed the plant’s managers would soon convert the entire facility to lean cells. (This, by the way, was my first awareness of the need for yokoten – the spread of a good idea, once demonstrated in one place, across an entire organization.)

In the following years I observed the Shingijitsu consultants from Japan, and many other advisory groups inspired by Shingi, creating model lines in many manufacturing firms to demonstrate the power of lean thinking to clients. These followed on in the tradition of the TSSC model lines and the original lean assembly line that Toyota managers created at NUMMI in California in 1984, which was designed to illustrate the power of lean thinking to General Motors managers in GM’s traditional assembly plants around the world.

I subsequently found the model line concept extended to a remarkable range of activities, including patient flow through primary care facilities and hospitals in healthcare, document processing in financial service firms, permit processing in governments, and applications writing in software firms. In my mind’s eye I envisioned a world where the transition from batches and queues to steady flow and leveled pull would soon be completed.

But instead I have gradually shifted from a model line enthusiast to a model line archeologist. I have often found in my company visits that the model line never spread from the original example, and that in many cases the model itself regressed, with the progressive addition of inventories between steps and even deconstruction to isolated villages. Perhaps the worst case is the sad assembly line for the Model 3 at Tesla (in the same NUMMI building) with its enormous re-work areas, a truly massive retreat from the NUMMI model line of John Shook’s era. Clearly there has been something missing in the lean conversion process.

This missing element turns out to be a complete lean management system, with line management thinking that supports the needs of the model line. As I have taken this insight on board in recent years, I’ve rethought the purpose of the model line.

Rather than simply demonstrating the power of lean thinking, and designed to solve a specific quality, cost or lead time problem, I now see the model line’s highest purpose is to serve as a splendid diagnostic tool that exposes all the problems in modern management systems. Let me give an example from recent experience.

Imagine a large factory assembling a range of product families on a number of assembly lines. And imagine that one of these lines is selected to become a model of lean thinking. In the current state there is a striking lack of repeatable standardized work, an absence of team leaders to support the front line associates, no line balancing of the sequence of machines being assembled (which creates real problems because the labor content for the most complex machine is twice that of the simplest machine), many parts that don’t fit properly due to poor design and that require fiddling by line workers, and very poor reliability of various pieces of automation and test equipment used on the line. As a result, productivity and quality suffer even when the line is able to make the number of machines demanded by customers everyday (usually with considerable overtime and a large number of machines in the rework are at final test.)

Further imagine that this firm introduces repeatable standardized work developed with the production associates, proposes a level schedule with products carefully sequenced to support the standardized work, introduces a team leader system to respond to front-line worker struggles within takt time, addresses part fit issues by prescreening parts to ensure they meet specifications, and repairs all equipment to support smooth flow to meet daily production needs without overtime. In addition, the management introduces visual management in each work station and a daily management walk to look at the visual management boards in each area.

This sounds like a valid model line and for a brief period it may appear to be successful, with lower defects, higher quality, and production of exactly the items requested by customers. But what happens when the line is turned over entirely to traditional line managers and the pre-existing management system reasserts itself?

The locked schedule to permit parts delivery in the right amounts at the right time is unlocked to accommodate “hot list” orders from sales. The parts system cannot respond to this, meaning the balanced sequence cannot be run, and so standard work begins to unravel as workers scramble to deal with imbalances, the off-standard parts reappear, the equipment begins to experience downtime, and team leaders can’t deal with disruptions in a timely manner. Problems (which are listed neatly on the right side of each production control board) pile up but can’t be addressed, and the performance of the line deteriorates steadily toward the original level. Good grief. I nearly screamed as I observed this happening over a period of a few months.

What was needed instead was an understanding of lean management to lock the schedule (which Toyota has always done ten days out from production) and to create a pull system all the way back to production of parts in the supply stream. The problem of off-standard by design parts needed to be resolved through collaboration with product and manufacturing engineering, and team leaders needed the training on methods for introducing temporary countermeasures to deal with problems as they emerge. Finally, a preventive maintenance system was needed to deal proactively with equipment problems and daily management walks needed to be redesigned to immediately address any deterioration in performance.

In short, a complete lean management system was needed to align sales, production control, product engineering, purchasing, and maintenance to support the needs of the final assembly line, where all the sins of modern management emerge for everyone to see (if anyone takes the time to look.)

So the model line in this case has served a critical purpose: to raise awareness of fundamental management problems. The question for the organization in question – and many more like it – is whether the evidence will lead to a fundamental rethink of management. Perhaps I will be able to answer this question in a future Lean Post.


P.S. Tesla is once again in the news. I have offered several times in recent writing for Planet Lean to visit Tesla for free to offer a bit of advice on how to create a model line. But I’ve gotten no response. I see today that Tesla is advertising for 400 new assemblers to staff the new plan for 24/7 operations on the Model 3 line in hopes of producing the long promised 5,000 vehicles per week. So: I’m prepared to do an assembler job for free in return for freedom to offer advice and write about it. At age 69 I would seem to be a bit old for the task. But there seems to be lots of down time at Tesla, so I think I can stand the pace until they act on my advice, transform their management thinking, and get the line to run steadily.


The views expressed in this post do not necessarily represent the views or policies of The Lean Enterprise Institute.
Search Posts:
Creating Continuous Flow / Making Materials Flow Set
By Rick Harris, Mike Rother, et al.
Lead With Respect
By Michael Ballé and Freddy Ballé
Was this post... Click all that apply
23 people say YES
25 people say YES
27 people say YES
16 people say YES
Related Posts
11 Comments | Post a Comment
Bob Emiliani April 19, 2018
1 Person AGREES with this comment

My experience with Shingijutsu consultants, dating back to 1994 (when I was a Business Unit Manager at Pratt & Whitney, and thereafter), is that they do not work to create a single model line. Instead, each sensei facilitates four or five kaizens in four or five separate areas to create flow in each one of those areas (and also solve combined cost, delivery, and quality problems). A few months later, they would return and facilitate kaizen in one or two of the previous areas, and three or four new ones. So, over time they were working to convert the entire facility from batch to flow, on the shop floor as well as in upstream (order processing) and downstream processes (distribution).

The backslide you cite is indeed due in part to the lack of an overall management system that supports flow. In particular, the finance department (which you left out) because they control the metrics that force batching in production and purchasing, unlocking (throttling) the schedule (along with materials management), and so on.

The exposure of fundamental problems in management created by "model lines" typically proves unconvincing to senior managers due to a combination of economic, social, political, and historical factors that reinforce the status quo. This is the subject of my most recent book; it thoroughly explains why the "fundamental rethink" rarely occurs.

I would be happy to mail you copies of the book for you and your team if you like.

Reply »

Ralf Lippold April 19, 2018
1 Person AGREES with this reply

Jim - Thanks a lot for your thorough description of what is in missmatch. Similar I experienced at BMW when they ramped up production in Leipzig from its start in 2005. What seemed to work pretty good in the beginning, giving the teams much authority to solve early on arising problems in production (at a time when production was scaling, given and rather stable workforce with established working relationships) changed quickly at a certain point.

Replication of Leipzig's success to other plants was about to happen, the visionary plant manager got other task, further improvement and head count reduction (cost reduction) was asked and the global financial crisis was about to happen soon. Besides, the minute production was not scaling anymore and most processes were rather stable, bureaucracy was settling again. Quick Kaizen or other improvement actions were not in favor, and “numbers began to rule“. 

Organizational management and “learned behavior“ of staff (whether employees on the front line, office or up the ranks to the top) is harder than concrete or the “proverbial“ water when one crashes on its surface not smart enough.

Understanding the inevitable dynamics within the organizstional living system and moving steadily forward as Taiichi Ohno did with support of his superiors is something rarely seen, and on the other hand very much in need.

PS.: Jim, I'd like to join you on the shopfloor at Tesla Motors in Fremont sharing the experiences for an amazing auto company that could be so much more than just another mass producer.

Reply »

Mark Graban April 19, 2018
2 People AGREE with this comment

"Clearly there has been something missing in the lean conversion process."

If you're referring to Tesla, it's not 'something missing,' it's a complete rejection of Toyota and Lean thinking there.

They didn't listen to Toyota when Toyota owned a part of Tesla. I'm told that relationship went badly.

Elon knows best. Of course he's not going to make an invitation or accept a visit.


Reply »

Mark Graban April 19, 2018
2 People AGREE with this comment

"I now see the model line’s highest purpose is to serve as a splendid diagnostic tool that exposes all the problems in modern management systems."

Well, management certainly doesn't want that! Expose the problems with the way we've been doing things? NO THANKS!


Reply »

Shahrukh A Irani April 19, 2018

Good morning Jim:

It is sad that you learned what you needed to know back when you co-wrote your book LEAN THINKING with Dan Jones.  If it was published way back in 1996, that means you and Dan must have been working on it earlier.  As this very welcome and eye-opening post indicates, both of you lacked the fundamental Industrial Engineering knowledge and failed to expand your literature review to the massive body of knowledge on Cellular Manufacturing and Group Technology in the UK, Germany and Russia!  Thanks to you both, we have been misled for 30+ years that it is management that makes any company successful.  No, sir, it is managers who began working in the gemba to start their careers and learned Indusrtial Engineering (or Production Engineering, as they call it in Europe) ***first***.  They learned by doing and then, as they moved up in years, they became managers.  Your post is too little too late!

Thank you though for LEAN THINKING because it opened my eyes that, if I did not transition from academia to industry and self-learn the Toyota-style IE that you both plagiarised, then I would been about as bad an Industrual Engineer as all you management consultants at LEI are.

Best wishes,


Reply »

Jessie Reyes April 19, 2018
4 People AGREE with this reply

While what you say Mr Irani is correct - growing up in the culture with hands-on work at the gemba etc is best leadership skills for Lean. But I have to disagree with your conclusion that this is the only solution. It is NOT a practical solution for companies hurting today and need help but has no leadership that has grown into lean. Tell me. How are they to get this solution you offer on-time to survive if it has to be grown? What would you do at Tesla if they called you? Tell them it's over because your leadership team did not "grow up" right?

Also, I find your comments/attack on Mr Womack disrespectful. There is a place for those like him that are expert observers, creators and communicators and folks like us who have grown up in lean through hands-on experience at gemba. Jointly both types make a perfect world for moving quality thinking to all levels and industries.  


Reply »

Jessie Reyes April 19, 2018

Mr. Womack. I totally agree with what you say and your "plan of action" to work at Tesla as an assembler is inspiring! I actually thought the same thing - but would they listen at all knowing the culture the owner has allowed to grow there? Most likely not.

When I saw the Tesla story and several pictures on CBS News the other night I was totally stunned! Right away I could tell they were not lean and wondered how they could not be in this day and age in the automobile industry. I'm sure had they been lean and were practioners of Lean Management, they would not need to drive up cost, lead-time and loose more money by hiring and training a 2nd and 3rd shift that is most likely not needed.

I have had the opportunity on more than several occasions to come into a company before they make such mistakes and find the key constraints within a day with VSM to being stopping the bleeding. Then we set in place the phyiscal Lean and Lean Management needed to sustain the changes.

See you on the Tesla line!

Reply »

Michael Ballé April 19, 2018
5 People AGREE with this comment


Thank you, fascinating and authentic piece.

I remember Freddy struggling with this as well back in the day. The model line work they did with Toyota was simply fantastic - and every one was in awe of it.

The thinking then was that if Valeo could reproduce the same on every cell, it would create lasting competitive advantage. This, of course, didn't happen. And yet it did.Because of all the low hanging fruit in the early nineties, applying basic learning from the model line did improve operations considerably and then... 

After a few years struggling with deploying the model line learning across the group - I believe Freddy and a McKinsey consultant were the first to come up with the ill-fated and unkillable concept of roadmaps to "implement"lean, Freddy realized two things:

First, the model line was not about fixing the process and replicating the solution, but about people development. His sensei needed the model line to progress to teach him the next step, and the assumption was (correctly) that Freddy's own improve knowledge would change his day to day decisions as Industrial VP, and from then on the company. This turned out to be even truer when Freddy became CEO of another automotive supplier. 

Freddy then realized that model lines were experiences for managers to click or not, much like NUMMI was at a much larger scale in the US: a place to develop great managers by giving them the right kind of experience in a semi-controlled environment in the terms you describe.

Second, but maybe just as important, Freddy realized that all the model line gains, 30% of laborproductivity, were paltry in the light of the engineering gains at the product renewal change, almost 30% reduction of the total cost - which Toyota split 15/15 with the supplier.

Model lines, he figured out, were not about finding solutions to fix the process, but training managers to change their outlook, understand the key principles of TPS and work more closely with the operators with genchi genbutsu and kaizen. By changing how management thinks, you change the company. The model line is the first place for managers to experience the changes required of them, and indeed participate to be the change.

Secondly, changing any cell into a model line has massive production engineering and product engineering impact as deeper and deeper quality and flexibility problems are solved, learning that, if you look for it, can be capitalized to radically improve the value of your product. Just-in-time, in a way, is a system to involve all suppiers in Value Analysis hoping for the potential of Value Engineering.

I'd be the first to agree (and argue for) that lean has to be seen as a full business system, but there is also a clear place for model lines, if they are used as space to develop people and to learn how to make better products :^)

Thank you again for a great Hansei piece!


Reply »

Eric Buehrens April 19, 2018
2 People AGREE with this comment

First of all, welcome Jim Womack back as a regular contributor to the Lean Post. His trenchant voice will only make us all better lean thinkers and practitioners and I look forward to his future columns.

I am very intrigued by the idea of seeing “the model line’s highest purpose is to serve as a splendid diagnostic tool that exposes all the problems in modern management systems”. I think this inverts our usual thinking about yokoten, doesn’t it? Instead of seeing the model line as a positive example of value creation that will spontaneously spread throughout the organization because of the demonstrable improvements it offers, we use the model line precisely because of the data that its introduction often produces about dysfunction in the management system. This data – the root cause analysis of why the existing management system often ignores or rejects the example of the model line – in this view then becomes part of the value of creating it – IF the lean practitioners in the organization can do hard internal work of bringing the organization’s leadership to greater consciousness and clarity about this dysfunction and how to address it.

I think this is a very powerful perspective, not least because it suggests to me quite a different task for the lean practitioner, including the Lean Enterprise Institute and our work with our many partner companies. It suggests that it’s not enough to build a model line, show its efficacy and the value it creates, and then mourn the organization’s failure to adopt lean practices more broadly. “They don’t get it”, we say. “They didn’t change their basic thinking”. Perhaps true, but we need to build the capacity as lean practitioners to probe the organizational resistance and to work with the data that resistance provides. If lean is a “socio-technical system for identifying weakness and making things better”, the resistance with which many model lines are met suggests to me that we need greater facility with understanding and intervening in the social system of the organization.

Reply »

Bob Emiliani April 19, 2018
1 Person AGREES with this reply

Unfortunately, what JPW proposes is neither powerful nor intriguing. Rather than inverting our thinking, it has been obvious for three decades that "the model line" exposes problems in management practice. This takes us back to 1987 when Shingijutsu kaizen consultants first arrived in the United States.

As I recall, value stream maps were supposed to help "lean practitioners in the organization... do [the] hard internal work of bringing the organization’s leadership to greater consciousness and clarity about this dysfunction and how to address it." But, that did not work, nor, generally, has other methods making leaders aware of the dysfunction. 

"...we need to build the capacity as lean practitioners to probe the organizational resistance and to work with the data that resistance provides." The resistance has been thoroughly described in my new book. I would be happy to mail you some copies. Let me know.


Reply »

Jim Lincoln April 28, 2018
2 People AGREE with this comment

Mr Womack, your ability to distill the complicated down to the simple is amazing. The model line as a diagnostic tool and teaching lab is what we should be about in the transformation.

With much respect, I would like to expand on your teaching to suggest that model lines should in fact be 2 lines - 1 in traditional setup, management, and disarray and the other as leaned and managed according to your prescription.

In this manner the “learn by doing” is supported by visually seeing and comparing the individual parts of both lines in real time. I have applied this approach and have had the satisfying experience of watching the learner realize the “ah hah!” moment of understanding.

Thanks for your teaching!

Reply »

Search Posts:
Creating Continuous Flow / Making Materials Flow Set
By Rick Harris, Mike Rother, et al.
Lead With Respect
By Michael Ballé and Freddy Ballé
A Lean Leap of Faith
Governance Matters
Lean Government