Column Archive: July 2012

Dear Gemba Coach,

I am working as a lean expert in a European multi-national with more than 25 manufacturing sites worldwide. We have started a continuous improvement process with dedicated lean experts at each site. Until now our focus was on lean training. We link projects to a site impact-priority matrix. Management wants the improvements projects to be linked to key performance indicators (KPIs) such as on-time delivery and cost of quality. And they want these KPIs identified, and that we report our progress on how the projects are impacting these KPIs on a monthly basis. Could you guide me on how this can be done?

Tough question, but it can be done. I know a CEO who is an experienced hand at lean turnaround and had worked with a sensei for many years before—but his experience should give you hope.

So how do we do it? First, let’s take a step back and think about lean programs in large, geographically dispersed companies. Let’s be clear about recognizing that broader goals at every level will have components of explicit business results, as well as organizational capabilities.

I believe that the core lesson of Toyota is to shift from JOB = WORK to JOB = WORK + KAIZEN. In other words, everyone, from the CEO to the operator, must factor in the education aspect (learning) of any job as well as the execution part (getting the job done. People at different levels will have a different mix of emphasis. At the CEO level, the lean job is mostly education (with some execution for Very Big Decisions). At the operator end, it’s mostly execution with some education (dojo training to standardized work and participation to kaizen events). The following chart might help:

Middle managers have the toughest challenge with this complex mix. Their responsibility in terms of lean means taking care of both the education system and the execution system. The big risk is to follow the path of least resistance and treat the lean program purely as an execution system, whereas it clearly is an education one.

How to Start

How did the CEO I mentioned handle this challenge? Let’s start with the execution system. The CEO started by defining five operational indicators that related directly to his P&L challenges (the following are an illustration, and not exactly how the company defined its KPIs, but close enough):

• Accidents: he believed strongly that people don’t come to work to get hurt and was his first priority (every accident across the world is reported to him immediately)
• Customer complaints: which reflected how customers felt about the product, and was a strong lever of sales
• On-time-delivery:  to measure service to customers, and again a strong lever of sales
• Days of inventory: a direct measure of cash
• Productivity: measured as direct heads per site. There is no really good measure of productivity, so the assumption was that each site had a more or less stable volume and that reducing the direct headcount while improving OTD should mean productivity improvement

On a monthly basis, the head of the lean program would track the historical evolution on each of these indicators against a target curve of reducing by half in eighteen months.

The important point on the execution front is that each plant manager was held directly accountable for the improvement of the plants according to these indicators, whether he or she chose to do lean or not. Every month, the lean office produced Pareto charts corresponding to each showing how plants stood respective to each other on the operational performance.

On the education front, the group lean manager created standard kaizen workshops corresponding to each indicator:

• Safety risk analysis (accidents)
• Quick response quality control from firewalls and red bins (customer complaints)
• Production plan leveling (OTD)
• Pull system and SMED (days of inventory)
• Single-piece-flow layout (productivity)

Each plant has a dedicated, full-time lean manager whose job was to run a program of workshops, at the steady rhythm of one workshop per month (some did more, some did less). The key aspect of the lean manager’s role was that he or she was in no way held accountable for the results on any indicator – this was clearly the plant’s management responsibility. The lean manager’s job was to run enough workshops (and run them well) to educate plant staff to work with operators to achieve the required results.

The trick here is to make sure that the “lean managers” are not implicitly tasked with getting results: their job is education, and they should be good at this, but that’s as far as it goes. Plant managers must achieve results.

The CEO reinforced this by making a commitment to visit every plant (about thirty across five continents) at least once a year. During these gemba visits, the CEO asked the plant manager (and not the lean manager) to show progress. Making sure the program of training was on track in the plant was the job of the Group lean manager, not the concern of the CEO.

Gemba Walking

The other specificity of this program was the CEO’s personal interest in what happened at operator level. During his plant visits, he would walk the gemba and ask to see how the operators were affected by the changes and how involved they were with the improvement actions.

On the other hand, the group stopped all classroom lean training. No one was expected to “talk lean” or to know about lean. Education occurred hands on, through the kaizen workshops, and the plants had to solve specific problems to improve the indicators. Through their gemba visits with the CEO’s sensei, the CEO and the Group lean manager established a list of key generic problems/generic solutions, which were progressively spread across the board.

Not all plants had results. A few never moved at all. A few showed spectacular improvements. But overall, the entire group of plants started moving, proving once more that lean results are obtained by the dynamic of lean, not simply “respecting standards” in a rigid way. The company went through two eighteen months cycles and was sold as it was beginning its third.

If I reflect back on this experience, I would say that you’re moving in the right direction, but beware of doing so without a sensei. Your CEO needs to understand why taking cost of quality as an indicator is a bad mistake as opposed to customer complaints (sales) and ppm (EBIT).

Remember that lean is a practice, not a theory and much lean know-how takes shape in the form of passing on a lean tradition from sensei to deshi. So, with so much at stake, it’s well worth finding a sensei that can help you clarify the high-level principles as well as guide you in the nuts-and-bolts specifics. If your CEO is not personally driving your efforts on the shop floor, it’s unlikely that a more KPI-focused program will have better results than the previous one.

Dear Gemba Coach,

I hear what you lean guys are saying about, learning, “kata” and the role of managers as teachers. But in my years at Toyota I was never taught much. I learned a lot, but what my manager would mostly do is check the details of what I did and ask a lot of questions. I felt that it was very controlling.

Thank you for your remark, which certainly echoes what other people who have worked for Toyota comment on. Not having worked there myself, I have no definite idea on how much control is involved in Toyota’s management culture, but, from the outside, I do suspect it’s quite an important component. Let me try to respond to your thought by investigating the relationship between teaching and controlling.

Let’s say that at the gemba, performance is the outcome of the “best” process being performed by the “best” people. Since nothing is ever perfect, the “best” process is sought by constant kaizen, and the “best” people are developed by on-going training. The virtuous circle sought by lean management is that on-the-job training fosters ideas for kaizen and process improvement, and kaizen efforts are key to learning and individual progress in understanding the nuances of the job and how to hone one’s skills. (The opposite is a vicious circle in which a lousy rigid process punishes and demotivates people, so they do their job badly which, in turn, worsens the process further and so on).

The central point is that performance is the outcome of a dynamic relationship between performing the job as best we know and improving the process. The best process in the world won’t deliver the results if the people aren’t skilled enough to handle it, and the best in class people can’t succeed in a broken process. You need both.

Everyone Matters

Let’s reflect more on the training aspect of this dynamic relationship. Most of our mental images about teaching are from our memories of school or the experience we have teaching our own kids: lessons have to be learned. Kids are supposed to “learn” as their full-time job and play the rest of the time. They don’t work, as we do, at repetitive actions that need to be executed for a specific result. They’re also a lot more tolerant and less set in their ways than adults.

Adults, on the other hand, typically have a job in which they need to successfully carry out a task, very often a repetitive one – even though maybe, being an operator on a production line is rather extreme in that respect. But your executive assistant (should you have one) will need to repetitively organize your diary, deal with people seeking to talk to you, organize your travelling, open the mail that gets to your office, plus all sorts of added ad hoc tasks. Even as an executive, succeed for the company as a whole to perform. Indeed, a key aspect of “respect” in the lean sense is recognizing that every single person in the company matters and contributes to overall performance. But where is the learning in that?

Contrarily to common wisdom adults can learn, will learn, and also enjoy it – but adult learning is very different from kids’ learning. Adults have a solid baggage of experience, busy lives with little time to spare, and, yes, admittedly, less flexible minds. Where do we start?

Start Here

First, what do adults learn? Adult learning is goal-oriented. Adults find it extremely hard to pick up an entirely new skill, and they seldom succeed at it (as my woeful attempts to pick up the guitar in middle-age have shown). On the other hand, adults will focus on a specific skill that they feel can help their job, and try hard to improve that – and even enjoy it. Consequently, as a manager, if you want your staff to learn on the job, you need somehow to define the purpose – the goal of the learning effort. Failing that, they’ll choose what they individually feel they’d like to learn, which is rarely what they need to learn. If you’re not aligned on this, much unfortunate misunderstandings and conflicts will later arise.

I was recently visiting an engineering department where the department head wanted to teach a younger engineer how to anticipate plastic deformation in the molding process (something to do with calculating the differential shrinkage in the cooling process). The engineer wanted to learn new functionalities in the design software. This conversation was not going well.

A lean approach to managing and teaching would have helped. The lean way to align intentions is to define with the staff an indicator that reflects the purpose of the job we’re trying to do as best as possible. This is never easy and should not be taken for granted. For example, counting customer complaints is not the same as counting bad parts at final inspection. Counting customer complaints is about learning to respond better to customer’s dissatisfaction, whereas counting rejects at final inspection is about improving the process’ right-first-time capability. Both are linked, of course, and both are most likely necessary, but the intent, and therefore the learning goal are very different. In the previous case, the engineering supervisor had not worked out with the engineer a specific target of scrapped parts for the project, and was having a hard time grabbing the guy’s attention.

Now, adults legitimately are interested in what they’re interested in: you can’t pick for them what should interest them, and chances are they’ll be interested in very different things from what you’d like them to be. As a result, frequent discussions about the indicator you’ve picked, how it behaves, why, and what people are doing about it are key to foster interest. Yes, this is certainly controlling behavior, and it can easily backfire if you fail to align their interests with yours. But, ultimately, sharing purpose is about constantly repeating the conversation about “How are customer complaints?” “What is the OTD?” “Are we reducing lead-time?” and so on. This will never be “acquired,” or “ingrained,” or “become part of the culture.” It’s a conversation that needs to happen, always.

When Practice Doesn’t Make Perfect

Second, how do adults learn? Mostly by practice. Practice, however does not mean mindlessly repeating the same task until you get it right. Practice means having a clear idea of the right outcome versus all the wrong ways of doing the job first. And secondly, having an understanding of what you’re trying to do in order to steer the practical action towards the desired result.

Learning occurs through the process of shuttling between abstract thinking (the toolkit of rules that can be applied in new situations) and data mining (dredging detailed information from our vast store of specific knowledge). As we grow older, we get worse at remembering all the details, but we get much better at seeing and using patterns. Essentially, learning comes from modeling a situation time and time again until the back-and-forth toggle between principle and specific generate new insight/competencies (both appear together).

Practice, as such, is hard to do without someone who micro-corrects the activity towards the relevant detail of where something goes wrong. On their own, people can repeat the same activity for years, stumbling on the very same problem without seeing the relevant detail. This is why musicians or athletes keep getting coaching even though they’re at the top of their game. The coach’s role is both to correct the micro-step (pull your wrist in) and to steer discussions on the high-level rules (hold a mental image of the chi flowing through the universe). This, again, is very controlling.

The engineering department head was trying to go through his junior’s calculation and point out hidden assumptions. The younger engineer felt he was being picked on unfairly and was so busy brewing internally, he wasn’t hearing anything much.

You Are a Role Model

Which leads us to the third question: why do people learn? Teaching through practice is in fact very controlling. This is tricky because, in order to learn, people also need to feel their efforts are self-directed and that they can learn in their own way at their own pace, for their own reasons.

Acquiring practice requires both drive and determination. You have to want to put yourself through a regimen of repeated attempts and constant criticism from your teacher. Why would you do that to yourself? And at work on top of it? The third component here is role modeling. In a work environment, in order to put up with the controlling dimension of teaching, employees need to accept (consciously or not) their manager as a role model. This is not an unrealistic expectation as most people compare themselves with the reference group of others who occupy the role to which they aspire, and so it’s not unrealistic to expect employees to look up to their manager for guidance in values and behavior. The difficulty is to translate this into acceptance of narrow teaching on specific skills.

In order to create the conditions for adult learning, the manager must ask himself or herself why his or her staff would accept detailed corrections about how they do things. In the case I was witnessing today, the young engineer clearly – rightly or wrongly (I don’t know them well enough to have an opinion) didn’t look up to his manager. As a result, he was simply being subservient in the conversation, and not taking much in: he wasn’t learning.

The lean management framework attempts to address this difficulty by stating explicitly that the frontline supervisor’s job is to give eight hours of standardized work to the team members. People are required to learn the standards, and, furthermore, the emphasis on work standards and kaizen creates a working environment where detailed investigation of work and confirmation is expected. But still, the question remains, what does it take for a manager to be accepted in this role by his subordinates. Technical competence? Empathy?

A Right to Succeed

Ultimately, acceptance hinges on intention. People will be tolerant of many faux pas if they are convinced that the manager truly means well and that he or she has the development of staff at heart, and is not simply squeezing the lemon to get all the juice. In the lean framework, this is the value that we try to convey through the various dimension of “respect”: every one counts, and every one has a right to succeed – which means that managers have to listen and work hard to take away all barriers to doing the job right (mainly muri, mura, and muda).

Ultimately, real teaching (as opposed to telling just so stories) is quite controlling: it’s about checking how the student is doing and then stopping and correcting him or her as they go so that they actually learn from their mistakes (saving it for the end of the process and dishing out all mistakes at once is a far less effective way of teaching). Furthermore, teaching also involves picking the topics people need to learn, not want to learn, which is controlling as well. In the end, the real difference is not the act of teaching or controlling itself, but how it’s perceived by the learner: is it for your benefit, or theirs?

Dear Gemba Coach,

Every one says that lean is not just about cost cutting. But I fail to see how lean techniques can influence the top line. How can lean help in this area?

I was on the gemba yesterday in a service company that has recently lost several large contracts in a bruising competitive bidding process. We’re still trying to figure out how customers can expect the required service at the offered price and believe they will be disappointed with the purveyors they’ve chosen on price alone, but that’s what happens when purchasing gets the upper hand at your customers. So although there is hope these customers will come back, in the mean time there is a huge gap in the sales budget.

The company’s CEO immediately reacted by identifying a group of customers not targeted before. This company services large companies, and is now looking to interest smaller, more independent outfits. Half a year into the pivot, we were looking at the sales figures and although the company has succeeded in avoiding a loss, the sales numbers are not growing nearly as fast as needed to recapture lost gain. This is a classic situation for using lean techniques.

Lean Sales Strategy

First, the CEO asked the sales team to visualize on a large chart the budgeted sales for the month, and then visualize by blocks of types of customers. In this way, yesterday they were discussing the gaps between budget and actual (some are more, some are less) and the immediate reasons for this. It was an interesting moment for me, personally as I’d never been part of a discussion using lean tools directly in a sales environment before. And fascinating, because after the usual initial reticence (sales is different, not everything can be predicted, etc.) they started looking more deeply at what they where doing and asking why?

It’s true that lean techniques are mostly concerned with engineering, production, and supply chain management, and there has been little written about applying lean to sales and marketing. Yet I have no doubt that a case can be made that kaizen applies everywhere. None the less, lean is concerned with increasing sales and, to some extent, this goal takes precedence. Toyota did not become the world’s largest automaker by reducing costs alone. In a mature market, every car sold is a car taken from the competition. Toyota succeeded in becoming number one by steadily increasing sales and market share. The question is: what is the lean strategy to increases sales sustainably.

A lean sales strategy hinges on customer satisfaction. The aim is to make our product a standard in our customer’s household or business. Our product should be the one that all other brands are compared against, so that the decision to buy a competitor’s product should be questioned. Since repeat sales are generally easier to realize (i.e. cost less per sale) than acquiring new customers, the lean strategy is to make sure that anyone who buys one, will continue to buy. In effect, the worth of a customer in these terms can be computed as the dollar value of the product or service multiplied by the number of times the person is likely to purchase the same product in their life-time.

Forgetting the First Step

In the previous company case, the CEO also started a war room where two questions were posted on the wall:

1. What is the key to winning the trust of small independent customers?
2. What level of delivery could be done at the bidding price for the large customers?

As many experiments were conducted over the months since the contract loss, some answers started to emerge – for instance, it appears that independent customers are very sensitive to initial response time – the time it takes to give them an answer, not the time it takes to actually do the job, which clearly large customers are not. On the other front, the management team realized that as large customers had traditionally looked for always more service, the offer had been enriched with many side services which could probably be pared down for a stripped-to-bare-bones service.

In short, we were having a discussion about value. Specify value is the first principle of Lean Thinking, Jim Womack and Dan Jones’ cornerstone book that launched the lean movement. Yet, because this is hard to do, most practitioners skip that first principle and move on to flow: aha, this we can do! Yet, value is what it is all about, and what supports sales.

Making our product the standard for our customers relies on both manufacturing and engineering. Not surprisingly, one function that emerges strongly in a lean company is manufacturing engineering as the arbiter between engineering designs and manufacturing capability.

Total Worth

In manufacturing terms, customer trust is gained by:

• Zero defects
• 100% on time delivery
• Regular price reduction or value increase (adding new features for the same price)
• Show of good will in any after sales service;

Many industrial managers will complain of the cost of getting it right 100% for the customers, which can involve special transports, extra 100% quality inspections and more. However, if we measure the customer by his or her total worth, we see that the price of 100% right is not that high relative to the customer’s worth; and as long as we don’t slap ourselves on the wrist by paying for anything less than 100% right, we’ll never fix our problems.

Truly putting the customer first in manufacturing goes a long way towards making the product a standard in the customer’s environment. However, it won’t help much if the product is not designed fit-for-use.

Specifying Value

How can we design products so that they become a standard for the customer? One way of looking at a product or service is that it solves a number of problems for the customer. As Marketing guru Ted Levitt once put it, “People don't buy a quarter-inch drill bit, they buy a quarter-inch hole.” One can therefore list the problems our product solves for the customer. A teapot will solve the problem of brewing the tea, solve the problem of letting the tea getting cold, and solve the problem of serving the tea. Many teapot designs don’t solve the problem of not burning yourself when you pick up the teapot and not spilling tea when you serve it.

In effect, the more problems you solve for the customer (including aesthetic sensibilities: a teapot can also solve the problem of looking decorative on the kitchen shelf), the greater the chances of creating a standard against which all other teapots will be compared. The temptation would them be to look at all the problems you can think of and try to solve them all at once, and thereby creating the “killer app” that is going to take all the market away.

The manufacturing side of the previous company had long focused on the value its products provided to its customers (the equipment being the basis for the service business). Being adept of lean thinking, the product business unit manager had first focused on customer quality and complaints. After cleaning up an impressive list of issues (and doubling sales in the process) the fundamental questions started to be addressed: what value the product truly provides to the customers? In the end, the team more-or-less settled on the idea that beyond all bells and whistles, the product delivered measurement and distribution. Kind of like saying that what makes a great car is essentially the engine and the gearbox (of course, you need the rest but that is where the customer difference lies). So, in the end, value was specified as measure, distribution, and cost of ownership.

The upshot is that the engineering team then focused on improving the measurement system to a point well beyond current market specs – which, once successful (difficult project, lots of mishaps, took too long, and so on), delivered a spectacular increase in sales. Over the entire period from (1) fixing quality issues relentlessly and (2) improving the “core” function, market share almost doubled in a mature, saturated market. The fascinating part from a lean point of view is that customers had never asked for this improvement. Marketing had not believed in its market impact until, grudgingly, it was impossible to deny after the fact. The gamble was made by engineers trying hard to figure out what customers really sought.

Delivering Value

In the process of seeking value, the engineering team had to resist marketing’s call for the new product that would take all the market by having all the right features. The hard issue was to learn what exactly customers would buy as opposed to what they said they would (in cognitive terms, customer’s “espoused theory” is often very different from their “theory-in-use)

The lean approach is not about gambling all on the next brilliant product, but learning to narrow down value and deliver it. Rather than make a huge gamble on coming up with a new product that will solve all customer problems at once, we can commit to a new product release at regular intervals – say a product renewal “takt time.” At each takt, we vow to solve more customer problems than were considered last time. In this way, we can progressively enrich the product – occasionally transforming it radically along the way (at some point, by solving point by point problems some “killer app idea might emerge.) In doing so we will progressively extend the list of problems the product solves for the customer but without ever making a huge intuitive leap into what the customer might want in the future;

This progression is consistent with the lean idea of innovation: solving today’s problems with tomorrow’s technology, rather than using today’s tech to try and solve tomorrow’s imagined problems. Furthermore, this approach to innovation also dovetails with manufacturing inasmuch as some problems we solve for the customer can have their root cause in our own manufacturing processes. Customer complaints are the most accurate source of potential value. By examining complaints at the customer’s Gemba and from the customer’s perspective, we can discover customer usages or preferences we had no idea about.

Question Product Strategy

In serious lean transformations, the first P&L line affected is sales. Delivering on time in full makes the backlog disappear (which increases sales mechanically) and halving customer complaints has a much quicker effect on reorder sales than people often anticipate. Cash is the second financial figure to show dramatic improvement as inventories go down (not at the expense of on-time-in-full delivery!). And finally costs come down as smarter solutions are found to sustain the lead-time reduction and eliminate waste.

The top line in the P&L continues to increase sustainably if customers start using our brand for all the similar products they use – they consider our product to be their standard. To do so, we have to constantly reduce the cost of ownership until the value-for-money ratio becomes unbeatable. Five out of ten vehicles Forbes estimates will make it past the 200,000 Miles are Toyota or Lexus (http://www.forbes.com/2010/12/08/most-reliable-cars-2010-business-autos-reliable-cars.html).

Not surprisingly, even in a year of horrid negative PR, three out of five Toyota and Lexus owners remain loyal to the brand when they buy a new vehicle, according to the J.D. Power and Associates Customer Retention Study released late last year. This is what fuels sales sustainably.

To achieve this we look at who we sell our products to predominantly, and we impose on ourselves a rhythm of product improvement by solving a few problems at a time for our customers. This might mean down to earth cleverness or dazzling (but validated) new technology. The point is that as we do that, the gaps with competitors increase because by the time they decide to catch up, they won’t have followed the same learning curve, and will have to pay for the increase performance or new features in hard cash.

Using lean to increase sales means taking a hard look at our product strategy and asking the question: “how can we improve products in order to sell far more to existing clients” rather than “What new product can I come up with to take the entire market?” The latter question involves working with both manufacturing and engineering to come up with a product that will establish itself as a standard for customers in terms of both conception and delivery. If you have no sales challenge for your lean effort, chances are you’re not putting customers first. In lean we expect sales to increase sustainably by building a basis for customers’ trust, and earning their loyalty – in terms of hard cash.

The challenge for the service operation is now to correctly specify the value of its offering. Surprising as it sounds, this is not an easy task and few companies are actually clear on this – yet it remains the first step of lean thinking.

How Can Our Back Office Sustain Lean?

Dear Gemba Coach,

The improvement program at the back office of the bank at which I am a lean facilitator is slipping. For the past several years, we have improved operations by visualizing daily performance, conducting daily briefs and teaching team leaders to solve problems. We have progressed visibly in quality and productivity. Recently however I am finding our lean progress harder to sustain. Teams that were doing great slip back and we seem to have to constantly regain lost ground. Is there a way to make lean more sustainable?

There is, and it involves the question of whether you are pulling—but let’s spend a moment framing this challenge.

Sustainability is and will always be an issue in lean because of the human factor. We not constructing a mechanical organization that runs like Charlie Chaplin’s clockwork factory (and God help us if we were!); rather, we’re coaching a football team. And as any fan knows, a team’s performance is, by nature, hard to sustain over time, because it depends on the players, the coach, and the circumstances. The same types of challenges apply at the gemba.

Our lean lenses can help us analyze this challenge with more rigor. Basically, we can identify three specific workplace diseases that produce waste and which eventually arrest lean progress. They are:

1. No confirmation: a failure to understand clearly the purpose of the job you’re asked to do. People need to be able to confirm bad from good (as opposed to blindly follow procedures) and to have some autonomy to make smart calls in different conditions.
2. Overburden: are you certain that workers are in good working conditions without having to deal with any unreasonable barriers such as too much work, too many interruptions, hard to handle software, unclear processes, arbitrary managers and so on?
3. Unlevelness: every day chaos as work is done in fits and starts, long batches and last minute panics, too many procedural changes, and all the various forms of instability that stop one from getting a grip and thinking clearly.

The lean approach towards this is not to create the perfect process, but to strive constantly to improve the conditions of the working process. This is an important distinction. Performance is not delivered by designing and then tightly controlling the perfect organization (sorry to pop your Taylorist balloon), but by generating positive dynamics through engaging in daily problem solving. People will naturally function better as a team even individually when they are in the habit of working together to solve problems. Solving problems together stirs a pioneer spirit in even the most administrative back office where the main task is data copy, as well as the craftsman’s interest for the job, no matter how narrow and repetitive. The point is that people will be interested in what they do – whatever it is – if you keep them interested.

And there’s the rub: how can we develop the kaizen spirit in a sustainable, generative, continuous way? How can we keep back office teams solving problems together day in day out, when they’re sitting in a basement doing apparently pointless repetitive jobs, such as checking credit applications or routine IT maintenance tasks? Experience shows there are three key ingredients to keep the PDCA wheel turning:

1. The manager’s attitude: time and time again we see that managers who are engaged with their workers produce greater lean success. If the frontline manager is genuinely interested in improving performance BY working with her teams to solve local problems and to involve every brain in the room in finding local knacks to make work easier, then people will continue to engage in problem solving and progress as a team. I’ve been visiting some companies for half a dozen years, and the lean teams that sustain success are always proud to show me how they’ve changed this or that to continue to improve their performance. And it comes down largely to whether the manager “gets it” or not.
2. Improvements must benefit the team: in some cases, I’ve met with managers who do “get it”, but have failed to get their departments to accept that change is a normal part of the job. These guys work very hard at problem-solving, but experience recurring slip-backs and need to be re-motivated again regularly. At first, I believed naively there was some fault in them, that they lacked some kind of leadership trait I could not see. But then I realized it doesn’t have to do much with them as managers, but with their higher ups. In some organizations, workers are moved around all the time and/or the pressure for productivity is relentless. As a result, whatever improvement made on a local station won’t benefit the person who had the idea because either this worker is pulled somewhere else or management will take all the benefit and never give anything back.
3. The team must also trust its own ability to solve problems. Some well-meaning managers solve all the “small issues” in order to focus the team on the large problems. This seldom ends well, as all the people learn from this is reliance on the manager. Motivation stems from success. In order to keep motivated, teams must regularly experience successes on their own terms. As a manager this means cherry picking problems so that they are both challenging (a bit) and solvable.

Unfortunately, many frontline bosses consider that their job is to tell people what to do when—and when. They use this limited power to (deliberately or not) reward the people they like with “plum” jobs and punish those they don’t with the nasty work no one likes. The problem is that nobody who works for these bosses cares about kaizen because it doesn’t change the fundamental unfairness of work.

Enter the pull system. Let’s go to a glamorous back office where people are supposed to check movie and TV production projects for approval for grants from a large foundation. At the gemba, the work isn’t so glitzy: a team of people work in crowded offices making sure the administrative part of the folder is correct before passing it on to the decision makers. They are organized by “specialties” and are asked to track their response time. Since work comes in irregularly, when response time of one person increases beyond the three-week target, the head of the office is supposed to go and help this worker solve problems. This often comes down to giving that extra file to someone else. Some times it’s the last in the queue, some other times it’s the tricky case that we’d all like to get rid of. There is nothing particularly bad in the way this office works, nor anything especially surprising. There are files stacked all over the place, and frequent people dramas and breakdowns, which is surprising because the department head is a caring gentle soul who genuinely tries to help his staff.

One day, the department head switches to pull. He creates bins with three vertical slots and a central waiting queue. The idea is that files are opened by his assistant and put in the central stack, and from then on into one of the checker’s “three slot” bin when there is an empty slot. At first, they all argue against this because they are so specialized no one could possibly do each other’s job. So they have a working session to isolate the really specialized cases, but they quickly agree that it could be interesting to look at other files as well. In two weeks time, the inventory is down to zero (which is problematic because it shows there is one head too many) and the extra time is used to focus on many improvement projects that had been put on the side-burner for lack of time.

What happened? Pull system magic. The pull system takes away from the manager the decision of what to do, so that he or she can focus on how to do it.

In a back office situation, the pull system also evens out the workload between employees, which, it turns out, is a huge benefit to them. Evening out the workloads liberates time to solve problems, problems that are now highlighted by the pull system (when someone’s bin gets clogged up). People now see that solving problems will help them improve their performance, not some abstract group performance (careful there: we don’t measure individual performance, only team, but now every person can see their contribution to the team).

One further overburden issue is that, fairly often, the best worker gets the hardest cases – and so is “punished” in terms of performance for her competence. The pull system will highlight this and help people deal with the situation better.

Finally, the pull system creates a working environment with strong incentives for the manager to keep solving problems – if he or she doesn’t the situation will quickly revert to what it was before and every one will notice.

The earliest paper I’ve found on lean is a 1977 piece by Sugimori et al called Toyota production system and Kanban system materialization of just-in-time and respect for human system. Quite a mouthful, but there you have it: JIT and respect go hand in hand. Over the years, I’ve found that a pull system tends to be the best (or least worst) answer to the sustainability question. And even so it doesn’t sustain itself by itself, since someone has to ratchet up the level of Just in time to lower the water in the lake, make the rocks appear and keep people interested in solving problems.

I haven’t come across one situation yet where pull is natural. Every one has their reasons to explain (honestly) that pull won’t work in their special case. With back-office work, I’ve found that visualizing a central queue of files and dispatching regularly to individual queues is a good place to start, but you’ll probably have to make your one up as you go. My father’s sensei’s advice was: “don’t interpret just-in-time at your convenience.” Start from there, and see where it gets you – my bet is that any pull system will make kaizen more sustainable than none.