Don’t Let Benchmarking Replace Your Own Process Engineering


Through the @LEIhealthcare twitter account, one source of autotweets are the new threads started in the Lean Enterprise Institute forums.

One new thread was started last week, titled “Sitter Ratios,” that caught my eye. Unfortunately, many of the Lean web forums and groups (including various Yahoo Groups and LinkedIn Groups) have a lot of discussion that involves people looking to copy others – their improvements, their processes, and (in this case) their staffing levels. We really have to be careful with that…

First off, the poster doesn't mean “babysitters” for home, they mean a hospital job that's described on this website as:

A  hospital sitter is an individual privately hired by the patient or his  family members to come and stay with the patient at the hospital. The hospital sitter then provides personal care that the patient may not receive otherwise. Hospital sitters are a valuable resource to many families that are unable to take time off to be in the hospital around the clock. Hospital sitters provide these families the peace of mind…

In one respect, the sitter might be a “workaround” — if nurses aren't available as often as needed, we could ask “why?” and do things to improve the system and support processes that would allow nurses to have more time with patients (using programs like the NHS Releasing Time to Care, based on Lean). Why are the patients not receiving certain patient care? That's another discussion…

One thing I've seen in hospitals is that there's a general lack of Industrial Engineering (aka Management Engineering, in healthcare) basics that would allow a department or manager to determine the right staffing levels based on inputs including  patient demand, quality and safety requirements, and that hospital's processes.

Staffing levels are too often based on budget numbers or benchmarks of other hospitals.

Does Toyota set its staffing levels based on benchmarks? No way. They understand their process well enough to know what the right staffing level should be. And hospitals can do the same. If Toyota has better processes and less waste than GM, than Toyota's factory should have a different (lower) staffing level. I'm sure even the individual Toyota factories determine their own staffing levels based on their own process (how does the San Antonio factory vary from the older Kentucky plant?).

Back to the LEI forum post, it asked:

Does anyone know of a benchmarked standard for sitters in a hospital environment? Looking for a ratio if available.

Obviously we're working on improving this regardless of if we're better or worse then a benchmark, but I was curious if anyone out there knows a standard. If the benchmark says we should be better than we are now, we'll use that as our crisis to help promote change.


I'm not suggested the original poster just wanted to copy… but we have to be careful with benchmarks. I'd rather focus on my own process and use time otherwise spent benchmarking on understanding and improving my own process.

I'd hope an organization can find a mandate for improvement other than “our staffing levels are higher than other hospitals that we benchmarked.” There needs to be a natural spirit of improvement in a Lean culture. I doubt benchmark numbers are going to be very motivating for hospital staff.

Again, if your hospital has a different layout and different processes than other hospitals, you might not be able to use some other hospital's staffing numbers.

My understanding is that Toyota spends very little time and energy on benchmarking others. Focus that energy on your own system, I'd say.

Looking back to my post about the interview with Masaaki Imai, look at the story he told about Taiichi Ohno:

… one time, he was the manager in charge of a machine shop, and he understood that in the machine shops typically one operator was assigned to one machine. But from the standpoint of people's efficiency, he noticed that each operator had a lot of idle time while the machine was working. So experimentally, he assigned two machines to one operator to see how that could be done.   Then, after very careful observation, he was able develop several methods that made this possible for all workers. To do this they needed to standardize the work procedure, and that's when he first introduced standard work. Then he increased the number of the machines to three, and in about six months one operator was taking care of four machines.

Ohno didn”t go benchmark other machine shops – he looked at his own process and engineered it. I'd argue hospitals can do the same thing. Once, I helped a hospital microbiology lab determine the right staffing levels based on workload data (including demand variation) and how long it generally took to do work (including the variation and without setting “quotas.”). We made the case that they needed to add a 5th “plate reader” and management listened to the data. Previously, it was just “an opinion” that the lab needed more people. Engineering and process data trumped opinion and gut feel.

When I tweeted about this last week, somebody raised the issue of learning from others. I agree that high level efforts like the Healthcare Value Leaders Network (an effort I'm involved) in is constructive sharing – organizations visit and talk with each other, sharing their lessons learned about Lean. We have a good culture in the Network of “learn, but don't copy” (this is an informal guideline). Our efforts aren't about copying the exact standardized work or copying staffing levels. I think it's a better form of “benchmarking” (we don't call it that) than following a bunch of benchmark data that might not have context.

What are your thoughts on benchmarking, and why?

  1. Absolutely worthless, don't waste your time
  2. Sort of helpful, but be careful in how you use it
  3. I couldn't live without benchmarking

What do you think? Please scroll down (or click) to post a comment. Or please share the post with your thoughts on LinkedIn – and follow me or connect with me there.

Did you like this post? Make sure you don't miss a post or podcast — Subscribe to get notified about posts via email daily or weekly.

Check out my latest book, The Mistakes That Make Us: Cultivating a Culture of Learning and Innovation:

Get New Posts Sent To You

Select list(s):
Previous articleTampa Bay Business Journal Features Lean at Moffitt Cancer Center
Next articleHoarders and Stashers? Ask Why…
Mark Graban
Mark Graban is an internationally-recognized consultant, author, and professional speaker, and podcaster with experience in healthcare, manufacturing, and startups. Mark's new book is The Mistakes That Make Us: Cultivating a Culture of Learning and Innovation. He is also the author of Measures of Success: React Less, Lead Better, Improve More, the Shingo Award-winning books Lean Hospitals and Healthcare Kaizen, and the anthology Practicing Lean. Mark is also a Senior Advisor to the technology company KaiNexus.


  1. Sorry I can’t help with benchmark numbers for ‘Sitters’ Mark but I agree 100% they should perhaps ask ‘why’ rather than just comparing values such as this .
    You really hit the nail on the head with your comments about benchmarking numbers: “There needs to be a natural spirit of improvement in a Lean culture. I doubt benchmark numbers are going to be very motivating for hospital staff.”

    In my experience working with the NHS and healthcare orgs in the UK benchmarking best practices can be of real value.

    Robert Camp (who wrote one of the earliest books on benchmarking in 1989) developed a very structured and thorough 12-stage approach to best practice benchmarking.

    More recently a number of NHS hospital trusts here in the UK have started to use online benchmarking tools such as – I find a team approach to the assessment is always best where each question is discussed and they reach a consensus.

  2. Benchmarking should always start with “what problem am I trying to solve that I can learn more about?” If you don’t have a purpose behind your benchmarking, then it is just industrial tourism.

    I don’t think it’s a big deal to know what the standard ratio is, especially with a fairly new concept like this, as we might not know how much opportunity is there. But the difference between knowing the number and using it as a target is where companies get into trouble.

    Coming back to another point you made, I don’t consider it a workaround at all. These people aren’t providing medical care. It is a difference service, a different value, and a different skill set. It wouldn’t be the best use of the skill set of the nurse to attend to every want of a patient, if there is another way to solve the same problem. Of course, how it’s executed makes all the difference.

    Jamie Flinchbaugh

    • @Jamie – yes, there are some things the sitter can do that the nurse doesn’t necessarily need to be doing. But I think one of the concerns would be the nurses thinking “well there’s a sitter” and then the nurse isn’t as attentive to patient needs. Even if a nurse gets water for a patient (something a tech or sitter could do), the nurse is more likely to notice changes in the patient’s condition in the course of bringing that water.

      If sitters supplement nursing care (i.e., nursing time at the bedside is the same), I think that’s good. But if the sitter becomes an excuse to spend more time sitting and charting at the nurses station, then that might not necessarily be good for patients.

      Then there’s the dynamic of you possibly hiring your own outside sitter and how the hospital staff reacts to that…

  3. I seem to recall that, back in the day, Toyota executives came to America and visited several industries, including supermarkets, and were impressed by what they say, leading to new ideas in their own production system.

    That is different than a Lean Coordinator getting benchmark data, but the original poster does mention using data as the crisis to get something underway. If your own executives won’t look outside on their own, maybe a carefully chosen benchmark will shake them from their lethargy.

    Our non-profit social service agency ranks best in our metro area for low administrative costs. I really wish I didn’t know that, because is makes it too easy to move slowly.

    Conclusion: I guess I vote #2. Be careful how you use benchmarks.

  4. I’d go with option #2 as well: use with caution.

    I think with respect to the sitters I would ask: Does the presence of the sitter improve patient outcomes? If so, should the hospital employ the sitters vice the families of the patients? Clearly there is a demand for sitters, but what needs are they fulfilling? Are they, as you considered, filling roles the nurses should be filling or are they providing something different? Are they simply replacing the presence of a family member? Can they? Are the sitters just piece of mind or are they a transformational part of the health care system? Have we studied their relationship to and impact on the system? If sitters are simply filling a gap in the system between the care provided by the hospital and the family, the ratio might just be a measure of how busy the typical family in that area may be, it might be reflective of a local cultural effect, or it might be showing an epidemic (enough that it is common knowledge in the area) lack of attention paid by the hospital staff. Where are sitters in the value stream? From the customer’s point of view it appears they are value added. What value are they adding?

  5. Obviously benchmarking as described is of very little value and can even be dangerous as well as expensive. On the other hand benchmarking can be a very valuable improvement approach for some. I look at benchmarking as a muturity issue. While Toyota may not rely on benchmarking you can bet Hyundai did as part of their rise from producing crappy third world cars to competing toe to toe with the Japanese. Would benchmarking have helped Yugo? Probably not. Does anyone believe Toyota doesn’t benchmark? Maybe more than you think – I guess there are more than a few Hyundais and BMWs torn apart in a back room somewhere.

    Applied properly, benchmarking is one way of providing an efficient pathway to getting in the game. Once you are in the game you can then afford the resources to get to the next level via something like lean. The negative aspect of benchmarking is that many healthcare organizations struggle to make improvements in spite of the facts, it does tend to dummy down the learning quotient, it can fool you into thinking that you have “arrived” even when you haven’t even started, and you will always come up short in terms what you could have achieved with a continuous focus.

  6. #2 with be very very careful. I would say that is the right answer but 90% of the people do it in a way that is damaging so remember you are likely to do more damage than good (unless you are much more aware of system thinking than most). So knowing that do #2 and learn about system thinking, variation, Standard Work Instructions, PDSA, going to the gemba…

    Basically you can get ideas from others – then adopt them to your system. And you can find out, sometimes, a solution you were complacent about or had become difficult for you to improve with your “blinders” now that you saw what someone else is doing you can see new ideas in your workplace more clearly.

  7. This is such an interesting discussion. I think I can contribute some insight. We provide “hospital sitters” for families and I think I can address a few of your questions above. Generally, we are replacing or supplementing a family role. Our Companions provide no medical care but are certainly integral in the emotional support and comfort care of the patient. Perhaps we are doing tasks an aid might do but we are dedicated to one patient. Although nurse ratio does vary from facility to facility, in our area, the nurses are so busy that they welcome our presence. However, we have found that our patients are NOT neglected because of our presence but actually receive better response rates when a Companion is present. In our area studies have shown that when a patient rings for a nurse, there is a 46% chance one will come on the first ring. When one of our patients with a Companion rings, a nurse arrives 95% on the first ring. We believe that our patients have better outcomes with the additional support and they also have a better impression of their hospital stay which is reflected in the patient satisfaction survey scores.

    • Kathleen – thanks for the comment and for the data.

      “Sitters” are a method and approach that could be used effectively or could be used to detrimental effects, I’m sure – sort of like bar coding or EMR systems. Some hospitals are going to “do it right” and others might end up creating problems.

      That’s a matter of the management system, not a negative reflection on the concept of sitters.

  8. At my current employer, we use “sitters” primarily for patients with delerium or dementia that result in behaviors that would negatively effect their health care (for example, a patient who is always trying to pull tubes out or get out of bed unsafely). The sitters then observe and report to nursing staff. In a better world. they would be trained to intervene and calm patients, provide distractions, etc. I think it is very sad to think that sitter might be needed to provide essential daily care such as oral and personal hygiene and feeding.

    With regards to the above blog, Mark identifies the concerns over whether one site’s “benchmark” is meaningful at another site. More often than not, it is apples and oranges, so a business case based on benchmarks can often be discounted and not very influencial. But, I completely agree with learning from “how” another site achieved it’s benchmark.

    Another concern is benchmarking comparison may simply raise the bar to be the “best” among the worst, without actually addessing the underlying performance of the system. You might be able to get your staffing level increased, but you may still have overworked, burned-out staff that are working inefficiently and making potentially dangerous errors. This is most likely to occur when there is a disconnect between work demands, work load performed, and the outcomes desired from that workload. As one nursing researcher pointed out, “don’t confuse product and productivity”.

    I do suggest that when it come to bedside clinical care, coming up with one’s own “benchmark” is often very difficult. Workload data does not necessary equate to “performance” data, particularly with professions that are involved in complex decision making, communication, and service delivery. The current buzz word in healthcare is “interdisciplinary collaborative practice” where the role overlap between professions is embraced … this creates complex interdependences that effect workload and performance. My own research and consulting is focused on helping hospital organizations address this issue to optimize patient outcomes, safety and quality of care, efficiency, while still maintaining staff morale and avoiding the negative impact of excessive workload such as sick time, staff injury, patient adverse events, and HR issues such as retention and recruitment issues.

    My opinion: #2 –

  9. Like all improvement I think benchmarking is poorly understood.

    Come back to Deming and Ohno. Would this benchmarking exercise show respect for people? If it does, and you’ve something genuinely to learn, then do it. For example, if people in your organisation are genuinely crying out to see how others do it, then let them.

    It’s waste to expect people to come up with their own change if learning from someone else would make it quicker.

    Second, benchmarking MUST be with DIFFERENT organisations. The pressure in my place is that people want to ‘benchmark’ with their peers. Other mediocre examples of the craft. Elephants racing elephants but the winner will be a horse.

    And you must benchmark ‘processes’ rather than organisations. It is not Toyota versus Hyundai, but small car manufacture and assembly in plant A versus another conceptually similar process but done differently at plant B eg DVD player assembly.

    The differences will be huge – that’s the point. The car assembler might never box and ship within minutes, but they might learn tips and tricks that will reduce from several hours to not quite so many hours, that they might not have thought for themselves.

    And finally, lean benchmarkers need to keep their eye on the customer and value. Does my customer care that I assembled cars quicker because I saw how DVD player assemblers do it? Will they pay for the improved quality or reduced time? I always remember the Morgan video where the consultant stuffed it up because he didn’t understand the customers were paying for a car to be made SLOWLY. A rare case, obviously, but never lose sight of it.

      • Definitely. Lean, or benchmarking should never be about copying. But learning how another organisation solved a problem, so you can apply your own brain to solving your problem.

        But it is waste to spend excess time ‘learning’ if you are trying to solve a problem from the ground up, when solutions exist. I’d hate someone to go to to trouble of inventing a cellphone for their mobile workforce, when such devices already exist.

        I spent some of my life in a hospital that took great pride in reinventing the wheel, when software or other solutions ‘out of the box’ were close enough to what is required.

        Would you have your developers developing survey tools online, when something like might be a ‘good enough’ solution? I’d rather that time was applied to things that really haven’t been done before, and for which there is a demonstrated need.

        I shouldn’t get started on this. My country has roughly 80% government funding of healthcare so as a taxpayer my hippocket nerve hurts whenever I walk through some hospitals and see what passes for process efficiency.

        • I think some of the analogies get a bit far away from the point I was originally making. Yes I’d use SurveyMonkey (and do) instead of inventing my own web service. But if I had a piece of software that was going to be critical to my business (or would create competitive advantage), I might get custom software.

          I’d certainly buy mobile devices, but I might not spend a lot of time benchmarking to see how other organizations use them.

          I think it’s a matter of balance. No need to reinvent everything, but we shouldn’t just expect easy answers, such as “how many people do I hire?” or “what’s my nursing unit layout?”

  10. Not sure why copying is inherently bad. Did I miss something in the Checklist Manifesto? Isn’t the goal to copy what has been deemed a best practice? At the same time the book makes the point that not everything can be reduced to a checklist. If my community hospital was having central line infections or four wrong site surgeries in an eight-week period I hope they would be copying pretty darn fast.

    • Hi Anonymous – I think there’s a big difference between copying a general practice and overly specific copying. The Checklist Manifesto emphasizes that people need to develop their own checklists. The specifics of a checklist might vary from hospital to hospital.

      I think it would be somewhat irresponsible and ineffective for a hospital to take a shortcut of just buying a checklist “off the shelf” instead of taking their own team through the process of developing a checklist that’s specific to their hospital.

      Asking “how many people should I have working in this area?” sounds like overly specific copying that shortcuts thinking. Learning from someone else about how they determined their own staffing levels is probably “good copying.”

  11. And isn’t not copying a form of waste? I think we copy all the time to great benefit. Aren’t the Toyota tools just best practices that have been codified over the years and every time someone does smed or kanban cards without doing a full blown RCA simply copying? Certainly copying all by itself isn’t sufficient but in the real world copying has its place. I’m guessing that Toyota wished it had copied Nissan’s automatic throttle shut down a few years earlier.

  12. […] Don’t Let Benchmarking Replace Your Own Process Engineering by Mark Graban – “One thing I’ve seen in hospitals is that there’s a general lack of Industrial Engineering (aka Management Engineering, in healthcare) basics that would allow a department or manager to determine the right staffing levels based on inputs including patient demand, quality and safety requirements, and that hospital’s processes.” […]

  13. At the end of the day, this is a debate about nothing. Copying, or benchmarking, have their place in the scheme of things.

    We teach at work “problem solving” – all the tools are there to help you solve problems. You should be able to build and staff a hospital ‘out of the box’ based on good working examples.

    Such a hospital will never be as good as a learning hospital, one that continuously improves. But to say you shouldn’t try as quickly as you can to move towards the practices of the acknowledged leaders is sour grapes.

    Toyota copied Henry Ford as far as that went. As far as using Ford as a base could logically be taken. They then embarked on a new path. Sometimes in my work I have to quote ‘the best’ just so people even realise it is possible to do something different from the way it is being done today. As Lean practitioners, we can forget some people just don’t realise they can be freed from the tyranny of the present.

    So I’d like to see a healthy attitude towards copying, immitating, be inspired by the way others work, up to its useful limit, and also people learning to solve unique problems themselves.

    Even Lean, to an extent, is just a series of concentric workarounds, each smaller than the previous. Like Russian dolls. We will never reach the central one, but hope each new one is smaller than the last.

  14. Benchmarking is not inherently bad. It is just misused. A few years back I watched some high paid consultants go into an industry leading company and share benchmarking data with the leaders of the company. The leaders then blindly took that data and regressed the company practices back to the mean.

    More recently I’ve seen benchmark data used out of context to make decisions that will be detrimental to the long term prosperity of a factory. The problem is that benchmark data is just that – data. No more no less. Benchmark data becomes useful to an organization when a person, team, and/or organization converts that data into a measurable improvement. Once again, the human mind is the most powerful element. Therefore our greatest investment should be in the development of our people to take data and convert it into improvement.


Please enter your comment!
Please enter your name here

This site uses Akismet to reduce spam. Learn how your comment data is processed.