When is Lean… Not Lean?

I have been thinking a lot lately about how the Toyota Production System was developed. Unlike those of us who have books, websites, and training programs in abundance, Toyota engineers took their process of assembly-line manufacture of automobiles and created, in incremental steps, the methodology that’s now known as Lean. It took shape over a long period, as different contributors added their ideas to create a strong “House of Quality” for the Toyota Motor Company.

But – what if these bright people had not come from automotive manufacturing? What if they had come from (for example) healthcare?

I know that this is akin to imagining what the earth would be like today if there were no moon. (Just think – no tides. Different air and water currents. Little shore erosion – no sand? Hard to imagine!)

  • What if their initial process studies were not based on repetitive motions that could be adapted to robotic or at least automated mechanisms?
  • What if their major source of variation was not the mechanical devices, but the people who provided the process?
  • What if their processes were not amenable to on-the-job training, but required differentiation of skill and ability at the level of advanced education and training?
  • What if their product was not someone who “ordered” a tangible object, but someone who showed up unannounced with a mysterious problem that couldn’t be solved by just looking at it?
Handpicked Content:   New Blog Alert: Six Sigma for Corporate Real Estate

It made me think that if those bright young engineers had worked in healthcare, that what we now call Lean would be vastly different than what was developed by Toyota.

And, as a related thought, it made me wonder what we are doing by applying lean manufacturing principles to healthcare.

Now, I’m not the healthcare equivalent to Eiji Toyoda. Or Shigeo Shingo. Or Genichi Taguchi. Or any of the other brilliant minds that helped to develop the Toyota Production System.

But, am I doing a disservice to my providers and customers, when I try to fit lean manufacturing methods to a highly technically-skilled service environment?

I’ve heard over and over againthat lean can be adapted to any process, anywhere, in any industry or branch of service. I’ve done manylean projects myself, and seen the very tangible benefits that value-stream thinking and creation of flow can produce, along with level loading and consideration of takt time.

But are we only seeing the tip of the iceberg? What further benefits might we see if we developed a “service system” that used tools uniquely intended for service processes, rather than adapting manufacturing tools as best we can?

Just asking!

Comments 15

  1. Jon Miller

    Hi Sue,

    TPS and lean aren’t systems suited specifically for manufacturing. Principles such as "eliminate waste, variability and overburden" or "pull instead of push" and "build quality in rather than inspect it in" and "standard enable improvement"are as applicable to complex services as to manufacturing.

    You may be confusing lean tools visible in manufacturing with the essence of lean itself, which comes from its guiding principles. A lean healthcare (or any service) system will come from those who understand these principles and pioneer their application in each specific environment.

    In terms of methods, "observe and document reality" is hard to beat, even if it does originate in scientific management and industrial engineering, both manufacturing-based. "Make rapid changes as experiments" is also almost purely the scientific method, though squarely lean, and shouldn’t be set aside because of daunting complexity of service processes such as healthcare.

    Lean probably got its start in repetitive automotive manufacturing because that industry provided the right business environment to evolve such a lean system. The healthcare business environment had no such need, so there may have been a "Taiichi Ohno of healthcare" but he would have gotten nowhere.

    The automotive industry offered repetitive work, complexity of product and process, the opportunity to redesign the work significantly due to regular model changes, and technology advancement and insertion over the decades.

    The business environment for healthcare is just barely getting to the point where a process innovator can make a lasting difference in building an efficient delivery system. Pressures from payers, the gap between the number of nurses and the growing patient base of baby boomers, soaring costs, the quality and safety concerns due to nosocomial infections and medical errors, medical tourism, etc. are all reasons for change that the healthcare industry was not facing 20 years ago.

    The universal principles of lean would have looked the same, regardless of which industry it grew up within.

  2. JConsidine

    The whole issue of takt in service industries is tricky – in many of the environments I’ve supported, it ranges from completely unpredictable to monthly or even annual demand for certain processes.

    I particularly encountered this while supporting a job shop, where customers’ own forecasts were generally useless. Meanwhile, corporate is leaning on management to get "leaner" with scheduling operators, while also squeezing OT and overall labor budget, and demanding an increase in overall "on-job" time. All while, by the way, guaranteeing very short turnaround,cost competitiveness and defect-free product.

    I think many of us in the non-manufacturing world would love to have a relatively stable takt to work to.

    That said, is it possible to successfully implement lean concepts in environments with no real takt? I’m thinking creative/consultative type professional environs, with a lot of rule by exception and not a lot of standard work?

  3. Rich Boehling


    It would appear you see what I’m getting at. If one can not define a "unit of value" other than time, how do we know if customer value is being acheived? In my opinion, the service side has are two basic types of transactions, 1) a deterministic type, processing a loan application for example, a definabable beginning and end 2) indeterministic type, the value of a therapist service or legal counsel. My sense is if the transaction is detreministic, then it can be measured and thus value can be assigned.

    A bank knows, for example, the value associated with a loan (if you can get one today), thus the loan processing transaction has value as part of a work breakdown structure or value stream of the entire loan. Thus a unit of value (or unit of measure) can be established which coorellates directly with labor cost. The challenge is to "mine for the units of measure" in a detreministic transaction which really do this.

    Returning to my thoughts of before, is it possible that in service side lean, the process that needs leaning is human work process. If so, then "what’s in it for me" becomes the paramount issue facing service side lean.

    As someone said before, what appears to be missing in american lean is the people respect pillar. For service side lean in particular, I think that may be the killer for Sue and her Health Care quest.

    More discussion would be appreciated.

  4. JConsidine

    Rich, the concept of mining for the right unit of value is critical – many shops really struggle with this, usually much to their consternation.

    Time as the unit of value often presents problems by itself unless it is properly defined.

    For example – we’ve all probably had the nightmare of going to an ER and having the entire experience last several hours, with only a few minutes of actual care.

    Most patients would say, faster is better. If you reduce faster is better to its absurd conclusion, you end up with zero cycle time.

    The reply might be, "take as long as you need to treat me properly, just don’t waste my time by keeping me waiting at different stages of the process, answering the same questions over and over, etc."

    Now we’re getting somewhere.

    An important point here is that services are not monolithic – there are truly transactional environments, like retail e-commerce, call centers, loan processing, car rental counters, etc., where the variation in the work is relatively smaller than others, like hospitals, engineering firms, other situations where there is a great deal of variation in the unit of value itself.

    This is why I posed the question earlier – in these less structured transactional environments, it’s very difficult to employ things like takt, heijunka, kanban, etc. From the Six Sigma side, it’s often difficult to reduce variation in the process when the bulk of it occurs within or because of humans.

  5. Rich Boehling

    Perhaps we should discuss the nature of the employer/employee contract. The employer provides a guaranteed wage to an employee and the employee offers a promise of effort in return. Ultimately, it is up to the employee to determine how much effort to give for the wage received. The variation you mention is related to the fulfillment of the employee’s promise of effort.

    Let’s go back to your ER example, what would happen if the ER processes were improved to deliever first quality through put only….the right patient gets the right care in the right amount with no waiting, with correct outcome (no rework). I think you would say… we can put processes in place but unless the ER staff participates fully with the program (keeps their promise of effort) then it will be to no avail. Because ER staff is paid based on time (input measure) not output it is likely this approach would fail. But should ER staff compensation be tied directly to "units of value" with clear line of sight to performance, then compensation (wage) is tied to process output being first quality throughput.

    Essentially, I’m suggesting the possibility that service side lean initiatives must include a feature that directly aligns comp to process success or simply put ….the pay system sets takt.

  6. Marty Yuzwa


    I too am in the service industry, and have spent time thinking of the same concepts. There are two main differences that I have found that I am trying to find and develop different tools to implement and train:

    1. There is usually minimal concept of Push systems or Over Production in the Service industry. Most processes are by definition Pull. So the "worst" waste of Over Production is mostly nonexistent. I tell people that this makes Lean even easier for Service processes to implement since they are already in pull mode. However, it makes it awkward when defining the seven wastes.

    2. In manufacturing the value created for the customer usually is evaluated by the end product after the end of the process. In the service industry this value is delivered multiple times throughout the process every time there is an interaction with the customer. It is therefore much more critical to define those customer contact points, and it is also more difficult to measure them. I have not found a good tool yet that is perfect to evaluate these value-add contact points, but Jim Womack’s new book, Lean Solutions did help.

  7. Rich Boehling

    Sue and Marty,

    The "contact points" for healthcare (at the hospital unit level )are pretty well defined by the US Army through accuity scores. Higher accuity patients require more contact, ie procedures. I don’t know about all of critical care, but a pilot project at an east coast hospital demonstrated an interesting relationship between patient outcome and accuity scores fulfillment for the ICU.

    Worth looking into.

  8. JConsidine

    One of the biggest problems I see with Lean is that the second pillar of lean is generally missing. This also holds true for most blog posts and comments about the subject as well.

    Respect for people.

    It seems to be profoundly lacking in American culture in general, and in business in particular. With a healthy understanding of respect for people, the rest of lean’s principles become much more intuitive, much easier to spread throughout an organization.

    But too often the focus of even us practitioners is reducing waste with an eye towards the bottom line, not because it actually demonstrates respect for the workforce by asking them to help in eliminating the things that not only add no value, but are detrimental to producing value.

    Of all the industries where respect for people should be overflowing is in healthcare – how much is really there? Among the front-line workforce, I would say quite a lot. Respect for people seems to become scarcer the further away from patients a staff member is.

    Moreover, Toyota is not just the TPS (they do not talk about "Lean") – there are 14 principles of management that are also critical to the success the company has produced over the years. How much a part of a typical lean journey are these? Probably not much at all.

    Once again, Lean practitioners focus on waste reduction to the exclusion of the rest of the system, which is why the results are just not there, regardless of the industry, and why Toyota is still an enigma to most western business leaders.

  9. Sue Kozlowski

    Hi Jon, I agree completely that lean principles can be used across many industries. And, that many of the principles are excellent for all settings. You’re right, it’s hard to beat "understand the current process" or "look for sources of waste to eliminate." My question had more to do with this: Are there ADDITIONAL lean tools or principles that might come out of healthcare, or other services, if we looked for them? — Since we don’t have the same business setting as the automotive industry. But your points are right on target for our current state in healthcare!

    James makes a great comment about respect for people. Part of my own "aha" experience was learning the lean principle of trusting the workers to know the process best. I had been brought up to believe that, as a manager, I should be the knowledge source. Ten years ago, I never would have trusted any of my entry-level employees to come up with a good improvement idea – they didn’t know the whole picture, they didn’t understand the budget issues, etc. Now I don’t make a move without the front-line workers being part of the team.

    I get a great response when I’m teaching about this, when I relate the concept that an organization should demonstrate respect for its workers by giving them only tasks that are value-added! And that the importance of "management" is directly proportional to the effort they put into making the front-line workers’ jobs easier to perform and to build in quality. It is a new concept for many of my healthcare audiences.

    Thanks to you both for sharing your insights!

  10. Rich Boehling

    Thought I might offer and additional perspective on "service side" lean. It have been my experience when the machine sets the pace of work, workers respond to the drum beat (to some degree). When no objective pacing is available we tend to rely on the individual to "set the pace". Perhaps the service side lean is encumbered by a lack of an objective drum beat. I would like to hear thoughts on how to set the pace when there is no objective tool to pace the work to acheive flow.

  11. Rich Boehling


    Perhaps we should focus on a specific process. I always use a Work Breakdown Structure to define units of value, sort of like value mapping but more akin to Earned Value Management. I’ve learned the unit of value must have three attributes 1. time bounding (no lag) 2. be discrete (measurable) 3. possess self evident quality. Each contact point with the customer or what I call a customer surrogate, must have these three attributes. The methodology expands the control volume until you have achieved these three attributes….then you set the unit of value for that piece of the process. The control volume may be large or small depending on the transactional nature of the service process.

    Send out a case and I’ll walk you and other readers through the method. The idea is to tie employee compensation to units of value and let the pay system establish takt.

    I hope I’m not talking Greek.

  12. Marty Yuzwa


    Sounds interesting (although I’m not in the healthcare field). I couldn’t really find much through google. If you come accross something, please send it on.

    The question I have is once you have identified these critical points by different customers, how can you measure the success of the interaction. Service employees can "go through the motions" and still not create the positive value-add interaction we are looking for.

    The best that I can do right now is to measure two different types of metrics:
    1. After-the-fact customer surveys. We use a Net Promotor Score survey. However, this is lagging, and doesn’t always tell us what part of the process broke down – unless we get really specific with some of the questions.

    2. Symptoms of poor customer interactions such as invoice disputes or rework that comes back to our shop, but again these are lagging in nature.

    Of course we can also measure process speed and
    "finishing when expected", which we know is important to customers, but this is only part of what defines value.

  13. Rich Boehling

    Marty and I had a short chat over the phone….here is an update.

    We discusssed the concept of "customer surrogate" being a unit of value which best reflects the desire of the customer. In the case of a truck repair, perhaps the measure could be "meanmiles" between failure(next repair) if the customer desires relability. Or another could be meantime between contacts, a customer might want an update every few hours, bigger repairs more contact.

    This presents an intesting parallel to Healthcare….a uroligist friend of mine told me that he has never been sued because he demonstrates care for the patient through direct contact, he believes as most in the industry that if a patient feels cared for then they will reciprocate with care, understanding and effort.

    I have seen the same in the ICU, high contact empathetic nurses (to both family and patient) tend to have better outcomes (no data just antecdotal). So maybe, a measure we seek (at least one) is the ratio of contact time to stay time. Acuity scores seem to indicate a causal relationship between the staffing level with patient task demand, is it possible that better outcomes can be acheived with better (more fequent) contact?

    Enough from me.

  14. Marty Yuzwa

    Sounds good Rich,

    Our process is a repair process for trucks. There are several critical customer contact points:

    1. Time of Initial Scheduling
    2. Greet Customer initially
    3. Customer Approval/Notification (only if customer requires approval or if job takes a turn we didn’t anticipate.)
    4. Customer contact for pickup.
    5. Greet Customer – hand over keys and final paperwork.

    I have a very detailed process map/value stream map that I can send you if you would like more specifics. What is your e-mail? My work phone is: 440-838-7308.

  15. Sue Kozlowski

    Wow, what a great discussion! As Rich points out, in many service sectors, although you can use queuing theory to some extent, there may be little predictability in the "demand" side of the equation. James helps us understand more about using takt time, and asks a great question about lean in creative/consultative activities. The conversation about "units of value" and linking pay with outcomes (rather than incomes) is fascinating. I particularly liked the comment about value to patients, to summarize, "if we can’t speed up the process, can we at least do every step correctly the first time, in an efficient manner?"

    Marty’s comment is very helpful about value-add contact points, and the difficulty when you have multiple service delivery points with the customer. For example, if I buy something, I typically have little to do with the actual manufacturing or delivery to the point of purchase. But in healthcare, as in other services, it’s the customer/patient who gets moved along the "production pipeline!"

    Acuity scores, as RIch points out, can be helpful when looking at the effort needed to provide the required services, and many hospitals use an acuity-based staffing system (except for those states, such as California, who mandate specific nurse/patient ratios – take a look at "California’s Nurse-Patient Ratio Law Saving Lives" ).

    Marty also makes a great point about lagging indicators – this is a conversation I’ve had very recently at my own place! The difficulty is that patient behavior and outcomes of treatment are NOT predictable. For any given treatment, the patient may a) not respond at all; b) respond as expected; c) have an adverse consequence due to an unexpected biological response (such as allergic reaction). Therefore the outcome, of effective treatment in a timely manner, becomes confusing when trying to process-map the care-delivery value stream!

    I very much appreciate all of your comments. Thank you for helping to clarify these important issues!

Leave a Reply