Who Sets the Standard?

Over the past few weeks I have had several conversations about standard work and who creates it. In each case I find myself trying to dispel one of the most common myths of the Toyota Production System: that the workers themselves create the standard work.” Most people who have studied or have been exposed to the notion of standard work will be familiar with this.

It has a nice ring to it, fitting with the notion of empowerment. It is a romantic notion, and a misconception. Many who teach lean are in love with this notion because it blunts the edge when delivering the message that lean requires a high degree of standardization, which to many means conformity, discipline and the loss of freedom. I have heard, and expect to continue to hear howls of protest by stating that lean in fact takes away the privilege from the worker of setting the standard. Dispelling this myth will be easier if we can understand and agree, dare I say standardize, on what we mean by “standard” and “standard work”.

To review the narrow definition of standard work, it is “the most effective method that that combines manpower, material and machinery”. Standard work is built on takt time, work sequence and standard work in process (SWIP). If there are any doubts on what we mean by standard work (also known as standardized work or standardised work) as part of the Toyota Production System, please refer to articles here, here , here and also here. Standard work is the cornerstone of the Toyota Production System and kaizen cannot exist without it. Taiich Ohno said that you must have standards, even if they are bad standards.

This type of standard work is based on repetitive work with a known customer demand. The takt time condition is not set by the person doing the work. That is determined by the customer who sets the customer demand and the management who set the hours of operation, or net available time. The pace of work is set by whoever designs the process, which may be the person doing the work but is most often an engineer or supervisor in the case of industrial work. Prior to applying lean and standardization, work is either self-paced or paced by automation, rarely customer-paced. Although there are arguments for allowing a method of pacing not strictly set to customer demand (mostly in service industries, creative and knowledge work areas is self-paced and which may not be well defined for creative work or in batch processing industries) at the least the pace of work should be synchronized with other internal processes to avoid waste, variability and overburden. Standard work requires that we do our work at a certain pace in a certain sequence, creating a limited amount of work in process. The worker does not set these conditions.

Standard work includes quality and safety key points and checks within each process. These are not set by the people who do the work. They are based on customer requirements, key quality characteristics, or regulatory and compliance issues requiring inspection and documentation of quality. Initial standards for processes are the result of a technical understanding of what it takes to repeatably perform a process safely, with good quality, on-time. Even in a lean operation only the most experienced worker has this knowledge and only as a result of years of experience, and even then they do not have the time during the typical work day to go about creating and documenting these standards.

While improvements can be made the method of how these checks are made and documented, resulting in quicker checks or more accurate checks, the person who does the work rarely has the option to remove these standards or even modify them. One exception to this is when the standard contains false or unnecessary checks. The worker may be able to identify that these checks are in fact no longer needed, or help redesign the process in such a way as to eliminate possibility of error. But this is saying that the worker can do kaizen. Since kaizen automatically implies that a new standard is generated, we should just say that “the person who does the work can kaizen their work” rather than “…creates standard work.”

The broad definition of standard work does not involve takt time and SWIP but nonetheless is governed by timing, a target completion time, frequency, and sequence. For example what we used to call “daily management” or simply “good operations management” that has gained new life these days as “leader standard work“. This involves visiting the visual boards to check target (standards) versus actual performance on a periodic (time-based) cycle. It is typically part of a larger daily, weekly, monthly cycle of management. While we can say that “the worker sets the standard” in this case, with the worker being the leader or manager, in fact where they go, when they go, and what they check are all determined by how the operation operates. The customer sets the pace of work, the management agenda set the performance targets for safety, quality, delivery, cost and so forth, and the start and stop times of each shift set the boundaries on when the leader performs their standard work. While it is not exacting in its specification, nor is leader standard work typically set by a 3rd party expert or specialist, about the only thing the leader can set in their standard work is when they make their gemba walks, and how exactly they inspire and motivate creative thinking and problem solving in the people they coach during the leader standard work process. Wouldn’t we rather have our leaders using their creativity in that area, instead of in creating the time-sequence-location standard?

Standards are conditions that must be achieved. Standards are set by experts, customers and regulatory authorities. Only in the case that the person who does the work is also the expert is it true that “the person who does the work sets the standard” and even then only when there are negligible customer or regulatory standards issues. Sometimes when we say “the workers themselves set the standard” we are talking about office work or non-repetitive professional work. I think in some cases it may be an attempt to appease the professionals who protest at having anyone else set a standard for the work that they do.

Just yesterday a medical doctor who is also a lean leader stated that 90% of the work a physician does can be standardized. That is great news. That may be one view and other physicians may howl in protest, but most of what we call “freedom” is simply unnecessary variation from the standard, due most often to the lack of documented standard, or a misunderstanding of what “standards” means. In fact most of us do not have control over the standards for the work we do, and only limited control over the creation of standard work itself. However we have almost unlimited control over the improvement of this standard, once we embrace standard work and kaizen.

All improvement must start and end with a documented standard. At the start, the standard may be unclear, not repeatable and unable to meet the targets. That is alright as long as it is recognized as the current best way and improvement follows from this point. We can say that “the person who does the work sets a provisional standard”. As long as we are following a repeatable method that delivers the basic customer requirements (standards) set forth, in the absence of properly documented standard work we should create our own. But setting the provisional standard is different from setting the initial standard based on key quality characteristics, equipment conditions and other working parameters.

Distinct from a provisional standard, the initial standard is created when a product or process is first launched. It is based on specifications from product, equipment conditions, and is typically the work of process engineers or their equivalent. In the worst case scenario these standards are never revisited, standard times are not met and conditions are not improved. Performance management through traditional accounting puts these losses in a variance bucket at the end of the day or the month, ask “who?” not “why?”, punishes and moves on. In the best case scenario, and in the lean way, these initial standards are immediately and thoroughly studied and improved until they reach a stable level performance. Thereafter these standards are continuously improved with input from the people who do the work. They may help with but are unlikely to be responsible for creating the standard work documents. Leaders maintain accountability for standards through visual management to make deviations and missed targets immediately evident within the hour, if not within the cycle, and by asking “why?” not “who?” and repeating the cycle of kaizen over and over again.

All standards are provisional, which is to say temporary. Even those given to us by customers and regulators are subject to change according to their whim. Standards set by physical laws are mostly set in stone, but as our processes and technologies change these too can be adjusted (although the underlying laws of the universe may remain).

Everyone can and should follow and improve standard work. Everyone can use their creativity and experience to think of better ways. But not everyone is qualified to take the initial customer requirements, convert these into product or process standards, and create initial standard work. Standard work is not created by the operators, however it is continuously improved by them.

So unless there are objections or evidence to the contrary or the assertions made above, let’s stop perpetuating the myth that “the person who does the work sets the standard” and that “standard work is created by the operator”. Both are misleading statements that cloud the important issue that most of us work in environments direly lacking in standards. We simply need to take responsibility and ask for standards so we can start improvements from there.

7 Comments

  1. Mike

    October 30, 2009 - 7:46 am

    The entire concept of standard work and takt time must be updated. One of my main disappointments with most otherwise excellent lean thought leaders is that they are still stuck in the 1980s with regard to their view of what modern manufacturing consists of. Most lean and standard work systems seem to be built around the concept of either “benchwork” or traditional assembly lines with much stratification and very specified work content. I maintain that most modern manufacturing of that sort is performed by automation these days wherever possible and affordable. Standard work and lean concepts really must be built around the idea of the operator now rather than the assembly line worker or work bench assembler. How you relate standard work and takt time to this concept has much more to do with lean in process industries than it does in traditional manufacturing. Really, it is time to move on from setting job standards that tell a person how fast they should assemble tab A into slot B in order meet the takt time, which hand to hold the widget in and which way to turn the screwdriver. Standard work has become MUCH more complex than that it if the lean educators don’t catch up with the realities of modern manufacturing they will risk being laid upon the same scrap heap as the batch manufacturers of the early 20th century.

  2. Rob

    October 30, 2009 - 8:53 am

    I think your view here is a little bit idealistic. The point of direct observations is that despite all the documentation and training for a given process, the operators do what they do. Oftentimes, they are so disassociated with the impact or potential impact (potential defects) that they don’t understand the worth of some of the steps. Once one operator stops doing it, others stop doing it. Or despite the effort, there is some inference or interpretation made by the operator. Many things can effect what really happens versus the original intent. Connecting key performance indicators with valued process steps in key to enforcement.

  3. Joe M

    October 30, 2009 - 12:33 pm

    I agree with Mike. I am a CI specialist in a low volume, high variation environment. I have always struggled with the “true form” of standard work. Our operators almost never make the same part twice in a row. One part has a cycle time of 1 minute and another can have a cycle time of one hour. This is a constant challenge not covered by traditional standard work combination documentation. However, I feel we have done a good job standardizing as much as we can (for now). I would love to see more “experts” post items that are not directly related to assembly line type production. As Mike said, Americas industry is not dominated by assembly lines at huge facilities, it is in fact found in small businesses with variation in mix and volume.

  4. Daniel Markovitz

    October 30, 2009 - 2:06 pm

    This is a thoughtful — and thought-provoking — post. But with your strict definition, it sounds as though there can be no standardized work for knowledge workers in creative and chaotic settings. How does SW fit into the production plan for an advertising copywriter, or the CFO of a company, or the HR manager?

  5. Jon Miller

    October 30, 2009 - 10:32 pm

    Mike – I think you may be mistaking standard work with work standards. We are not talking about job setting standards – and that is not standard work, the job setting times are work standards. I agree that there is less and less labor-intensive manufacturing in the USA but there is still a lot of repetitive distribution, fulfillment, production and service work that can benefit from standard work. Or if your specialty is automation and mine is not, we may simply be seeing two different parts of the same manufacturing world. Within manufacturing today standard work is alive and well. Assembly lines are still the primary means of putting durable goods together around the world and standard work is the best way of organizing this work. Try to visit a Toyota factory if you can.
    Rob – Lean is by definition idealistic. It is the endless pursuit of waste elimination. It is the endless cycle of setting standards, improving, setting standards, improving… Your point that upon observation people can be seen to not follow the standard is the VERY reason that we need visible standards – enabling abnormality management and therefore kaizen.
    Joe – I take your point, and would say that for high mix, low volume operations the repetitive cycle mass production standard work is not the appropriate solution. However just as in the example I gave for leader standard work, would we not benefit from having visible standards for jobs that take 1 minute and also for those that take 5 minutes? As Rob pointed out, what you THINK takes 1 minute may take 3 min if you have no standard to compare against. We don’t need to insist on applying the narrow definition of standard work in these cases, but we do need to set standards whenever possible in order establish a baseline for improvement.
    Dan – I gave the strict definition and then the definition of “standard” as well as a broader definition and how it applied to things like leader standard work – the work of a factory manager making their rounds for instance. At the very least we need to follow standards in the broadest definition – conditions set by the regulators, the customer, the workplace. Then there are performance standards even for creative work in terms of output per period of time. It can be advertising copy created, contracts negotiated, widgets sold, profitable ideas thought up. In the narrow definition it may not be possible or appropriate to talk about standardized work for knowledge workers because there too much variation in demand and work content to establish takt and work sequence. However on a longer time cycle we can still set the sequence that knowledge workers follow. If we map the process from acquiring a lead to closing a sale, there are a finite number of steps, and if you follow the Sandler method for example, it gets to be very close to standard work.
    All great comments that raise important questions about work and how to improve it. Please keep them coming.

  6. John Santomer

    October 31, 2009 - 7:01 am

    Dear Jon,
    For a standard procedure on document processing – if a standard procedure does exist that can handle an anomaly or abnormal case is encountered. One that the standard procedure does not have a clear outline for handling; can the person who does the actual “legwork”, “blazing the trails” and makes the new steps not stated in the standard procedure DRAW the standard for the anomaly or new case? The specialists and the lean leaders who were supposed to set the standards to cope with the abnormal case does not initiae any steps to place a procedure to handle the abnormality and people were told to still follow the “normal” procedures.
    In the absence of a clear outlined step to handle the anomaly, when asked why the procedure took longer, the people who did the “legwork”, who were obliged to find ways to bridge the “gap” were even BLAMED for not being able to complete the requirements quickly, in time which were not correctly or clearly advised to them in the first place.
    Can the same people DRAW the standard as the specialists and lean leaders refuse to accept the responsibility to address the anomaly?

  7. Rob

    November 2, 2009 - 7:56 am

    I think the main issue with standard work is that I all too often hear people confusing it with work standards or even engineering, quality or some other type of technical specification which has been imposed. The two terms become interchangeable in everyday language but they mean different things.
    Work standards tend to come from new product or process development and attempt to try to lock in certain work standards, for example, speed of the machine should be X, torque value should be Y, or drying time should be Z. In my experience, these are often backed-up by a change management process to allow controlled changes to be made, often as part of an ISO type of quality system. This is generally a reactive way of working, perhaps in response to a customer complaint or an incomplete or ineffective error-proofing device which allows too many rejects to be produced. A more proactive way is, as you said:
    “the person who does the work can kaizen their work”.
    John Shook said as much in one of his recent blog posts:
    “integrate people with process (the social with the technical). Nowhere does that come together more than in the form of standardized work and kaizen.”
    Standard work uses takt time (customer driven timing and process constraints), work sequence (the best sequence of building the product) and SWIP (how much in-process work, data, information, material, etc is allowed).
    For me, work standards come first then standard work then PDCA. Not the other way round.
    In addition, as stated in your comments, manufacturing is shifting away from high-volume, low-mix production to high-mix, low-volume production, as consumers increasingly want products specific to their exact needs. The key to success here is to learn how to schedule production effectively, avoiding bottlenecks that stop upstream processes and kill downstream ones. So, split each order into batches which can all be assembled in approximately the same amount of time by understanding how much time is needed to complete each step for each product. This is where SMED helps enormously.
    Finally, I have a question concerning the term “leader standard work”. Is this similar or the same as Layered Audits?