Home > Knowledge Center> eLetters> John Shook's eLetters> A Lean "Teachable Moment": Starbucks in The Wall Street Journal

A Lean "Teachable Moment": Starbucks in The Wall Street Journal

John Shook
Permalink   |   20 Comments   |   Post a Comment   |  
  |   RSS
I like to keep my consulting relationships quiet. Nothing like a provocative piece on the front page of The Wall Street Journal to blow any quietude out of the water.

This week's Journal featured an article titled: "Latest Starbucks Buzzword: 'Lean' Japanese Techniques." The article referenced me as the 'former Toyota executive who has been advising Starbucks on lean methods." You can also find a free version of the article here.

The blogosphere – the Journal on-line comments space and other websites – is buzzing. Starbucks is a huge flashpoint for bloggers anyway. Many folks hate and many folks love Starbucks. Many folks seem to hold both feelings about the company. Some seem to love hating them.

I won't comment on everything in the article nor will I reply to all the things being said in the blogosphere. But, there are a few matters that call for being set straight, key among them the charges (we've heard this before, from industry after industry as it encounters lean) that lean is nothing more than an efficiency campaign that is indistinguishable from Taylor's scientific management; and, that Starbucks baristas will become robots as Starbucks seeks to equal fast food joints in labor efficiency.

Of course, that is all far from the truth.

The problem with Taylor's Scientific Management: Who is the scientist when it comes to process improvement? Scientists must see real work to do science on the work.

Previously in this space, we've discussed the unscientific nature of most interpretations of Taylor’s scientific management. Those misconceptions seep into the dialogue of even very experienced lean practitioners.

There is a technical side of lean that seeks to make production – all production, all services, all work – flow from beginning to end as efficiently and effectively as possible. Traditional industrial engineering practices – including the ideas of Frederick Taylor – play a critical role in those aspects of lean.

Lean Thinkers (from Toyota Production System developer Taiichi Ohno to MIT lean production research team leader Jim Womack) credit Henry Ford with first establishing flow production, a precursor to the lean enterprise. Henry combined interchangeable parts with the flowing assembly line and showed it to the world. Production hasn’t been the same since.

Toyota then evolved Ford's flow production in two critical ways. First, technically, Toyota figured out (with great benefit for the customer) how to achieve flow production in lower volume, high-variety environments. Ford's flow production worked best when offering only "any color you want as long as it’s black." But, customers eventually demanded more variety. And when faced with the need to respond, Toyota showed us that flow is possible even in complex product mix environments.

Toyota's most radical innovation.

Toyota's second, and most radical, innovation was to answer the central problem that came with Taylor’s Scientific Management: the inhumane treatment of workers doing manual labor. Toyota revolutionized the technical side of lean production with the inclusion of product diversity into the production flow. But more importantly, Toyota revolutionized the social dimension of work, respecting workers brains as well as their hands. So factory workers become knowledge workers.

Toyota combined old IE Scientific Management principles and techniques with social dimensions appropriate for the modern world. Even workers who do "manual labor" with their hands are knowledge workers. Front-line employees become the scientists.

By redefining roles, Toyota changed the answer to the question of who is the scientist in scientific management.

The march of civilization is usually considered to progress from the Agrarian Age (with the Agricultural Revolution) to the Industrial Age (with the Industrial Revolution) to the Information Age (led by the knowledge worker).

So, Starbucks...

It seems that in early stages of embracing lean or any process improvement methodology, practitioners inevitably misunderstand and go through a period of neglecting the social dimension of lean. They try to determine the One Best Way to do work and then deploy (roll out) that Way, in a programmatic way, seeking compliance. What is especially interesting in the Starbucks case is that they can't even HOPE to do that, not with 10,000 stores in North America managed by kids managing part-timers to provide a high-end product and service. So Starbucks is working out a way that could be revolutionary. It leads to a better way of working for baristas that brings even better service for customers. And, oh, by the way, tremendous cost savings can come along, as well.

The way Starbucks has chosen to incorporate this new way of working is revolutionary. Traditional companies too often try to implement lean (or perhaps traditional industrial engineering, or six sigma or process reengineering), in a programmatic way. They do it that programmatic way because … simply because they can. Starbucks, on the other hand, couldn’t approach change on its massive, diverse level in a programmatic, straight-forward, cookie-cutter way. Not and provide the kind of unique, customer-oriented service they want to provide. They had to do it a different way. They have no choice but to do it the right way – through involving the people who do the work.

The comparison with McDonalds is erroneous and misleading. McDonalds very business model seeks a highly cookie cutter approach. Therefore, McDonald's may be successful in implementing traditional Industrial Engineering (Taylorism and all that – not lean) in a very traditional, top-down, programmatic way.

Starbucks decided long ago – and reconfirms this every day – that a cookie-cutter store approach is not the pathway to success for their product, which is a higher-end, higher-priced coffee that emphasizes the customer’s experience. (That is Starbucks’ explicit aspiration. Many people hate Starbucks. Some hate it because it is too upscale, over-priced, and pretentious. Others hate the taste of the coffee, complaining that it is too strong, too weak, too "burnt." Some call Starbucks loyalists with their particular espresso demands "coffee snobs" while there are also extreme coffee connoisseurs who consider Starbucks coffee to be undrinkable swill. You may or may not appreciate Starbucks' aspirations or agree or disagree that it meets them – that is not the point of this discussion.) Each Starbucks store is different. The footprint is different, the customer experience is different. I believe Starbucks wants the customer experience from store to store to be consistent but unique. McDonald’s wants the customer experience to be exactly the same, totally common from store to store.

Starbucks wants the customer to enjoy the experience of being in the store, of interacting with the barista, of hearing the barista call his or her name. Starbucks wants the customer to appreciate the fact that the barista is highly skilled at crafting each drink to perfection and to the customer’s satisfaction.

In each of Starbucks 15,000 or so stores (the gemba or place of work for over 100,000 employees), the next customer to walk in the door may order any of over 80,000 drinks from the nearly infinite available combinations. And then there are those custom drinks, with "quad shots" of espresso (popular with students during final exam week).

Far from becoming robots, think of the best bartender or waiter/waitress you've ever seen. Remember marveling at how he or she could handle orders coming from all directions, without missing a beat. That’s what Starbucks wants from its lean initiative.

Instead of barista's having to stop to search for things that are in the wrong place, or aren't there at all, the goal is to make as many things as possible routine so that the barista can spend just a few more seconds talking with the customer. That's the goal. No workarounds due to the line backing up, no short-cuts to get caught up – handling each unique order as it should be handled, in stride, without burden, and to the customer’s satisfaction.

No doubt McDonald's wants its customers to be just as happy, but they want to achieve that by making every experience exactly the same. Therefore, there is no problem with designing the work (with good Industrial Engineering built-in) at headquarters and then rolling that work design out to the masses. In other words, I would argue that McDonald's and the others aren't doing lean, but Industrial Engineering. The technical side of lean without the social side isn’t lean at all.

Starbucks is approaching lean with the intent of providing their baristas with the skills to do better work design on their own, as they go along. This is in total contrast to the uninformed charge that baristas are being made into "robots." If that is what Starbucks wanted, there are easier means to get there than their chosen method of introducing the concepts to each store and asking the baristas to work on their own unique solutions suitable to their own unique situations.

Starbucks and beyond . . .

By the way, the lean transformation Starbucks is pursuing is possible for all service and retail industries. Many service industries – especially healthcare – are discovering the power of lean. But retail is still a state of nature - the way employees have to stock and restock goods, often having the wrong items in stock is a huge opportunity for retailers everywhere. Did you know that, on average, grocery shoppers fill their baskets with what they want only a little over 50% of the time?! Did you realize that a good 20% of the time, your shoe store doesn't have the style or size you need?! Do you remember the last time (this morning for me) you received service that made you want to scream or just roll your eyes?

Back to The Wall Street Journal, all this publicity is probably unfortunate in the sense that Starbucks is still very early in its lean journey. Starbucks is approaching lean the only way it can, and has been very slow and methodical in developing it in a way that fits its Starbucks culture. It is fantastic that as a result of all this Starbucks may be able to provide a great model for other service and retail companies to learn from. But that's a big maybe (underlined, italicized, bold) – it's way too early to tell about that. I’m happy that they're trying and learning.


John Shook, Senior Advisor
lean enterprise Institute, Inc.

Summary of evolution of lean – from Chicago meat packing to Starbucks

Chicago Meat Packing -> Henry Ford -> Toyota -> Starbucks

Chicago Meat Packing and the technical side of lean production -> the disassembly line

Henry Ford and the technical side of lean production -> the assembly line

Toyota -> revolutionized the technical side of lean production with inclusion of product complexity (for customer benefit). More importantly it revolutionized the social dimension, respecting workers brains as well as their hands (so factory workers become knowledge workers, the scientists).
20 Comments | Post a Comment
Jon Miller August 7, 2009
Hi John

Thanks for adding these details to the WSJ article.

It's always unfortunate when a company invites journalists in to see their lean efforts in an attempt to curry favor with the Wall Street analysts, so it falls to us bloggers to provide depth and context.

I will look forward to seeing signs of true kaizen culture and employee engagement during future Starbucks visits.
Business901 August 7, 2009
I not only enjoyed the particulars of the situation but also the great insights on the roots of Lean- Ford to Toyota and "The technical side of lean without the social side isn’t lean at all."

Dispels many of the Myths in 2 sentences.

Thanks for the insights!
Andy Satt August 7, 2009
The key message for me in this blog is the clear statement that "knowledge workers" are the key to success. This is it!

In our world, where nearly everything is becoming a commodity ("What price?"), the only differentiation are the ideas and the brains of our associates.

Great to see that written out by John Shook.
Norm August 7, 2009
Awesome discussion on the evolution of lean and why cookie cutter approaches just do not work.

Thanks John - very inspiring

James August 8, 2009
Too bad more companies do not understand grass roots CIP.
James August 8, 2009
Too bad more companies do not understand grass roots CIP. Failing to do so put GM and Chrylser under.
Brian Buck August 8, 2009
John, thank you for this post. You have made me really excited to see how Starbucks becomes Lean.

Being a past partner, baristas frequently complained about how corporate engineered their individual store. I am very happy to hear their strategy to look at the barista's own unique solutions.
Liz Guthridge, The Lean Communicator August 8, 2009
John, thanks for sharing your insights and experiences.

From my perspective, Starbucks is an interesting case study because the company started on its lean journey several years ago in headquarters. Marketing Communications was one of the early adopters and was able to simplify, speed up and reduce sigificant waste in their marketing materials. I enjoyed hearing about their experiences at the first lean transformation summit LEI held in 2007.

Starbucks didn't use a cookie cutter approach then, and they're not using it now with the stores either.

Here's wishing them more success and here's hoping others will realize the value. Lean is a triple win--for customers, employees and the organization.
John August 10, 2009
Thanks to all for the great comments!

Thanks for rightly highlighting the role of bloggers in cases like this, Jon. In this particular instance, though, Starbucks didn't "invite WSJ in" to curry favor. This was initiated by the Journal. I think Starbucks agrees with me that the publicity is premature. It's better to get solid results first before rolling out the public communications machine.

"The technical side of lean without the social dimension isn't lean at all" does say a lot. My summary of lean history was indeed a summary, so omitted much. I also may have seemed to trash Taylor, Ford and McDonalds more than I intended. But, I think that it's undeniably true that, at the end of the day, you have to have both the technical and social dimensions in equal measure.

Right, Liz, about Starbucks starting their learning in some headquarters office processes. Seems that every company's path is unique to some degree. The question of "where to start?" bothers people a lot more than it should. It's not WHERE you start but HOW you start.

One interesting thing (another of the many) about Starbucks is that we can all go visit a Starbucks gemba every day. So, you will be able to see for yourself how they do as time passes.

john s
mak August 10, 2009
Hi, John, colleagues.
My name is Mikhail Kalinin and I'm branch and plant manager at 3M Russia.
I am in the business for more than 20 years and I used different concepts and instruments to improve effectiveness and efficiency of the work, among them: KPI, 6 sigma, BSC, Lean.
I found that it does not matter how the concept used is called. The key to success is the Leadership!
What do you think?
Thank you.
Frank August 10, 2009
These are very interesting parallels between the Lean approach at Starbucks and the emerging discipline of service design.

Thanks John for your thoughts on transforming lean concepts into the service sector. I'd like to hear your perspective on the concept of standardization in this context. You wrote:
"Starbucks is approaching lean with the intent of providing their baristas with the skills to do better work design on their own, as they go along."
Do you see a contradiction between this approach and the lean concept of standardization? It would be interesting to know, to what extent individual improvements in work design are being standardized across the organization.
John August 10, 2009
Thanks, Frank, for the observation and great question. It’s a question that comes up a lot. I am convinced that confusion around this issue holds a lot of organizations back. Part of the answer relates to my recent Column about the Suggestion System. More directly to the point is the fact that it is critical to distinguish between standardization and commonization.

What is the purpose of standardization? From Henry Ford to Shigeo Shingo to your neighborhood lean consultant, the answer you will hear is “to form a baseline for improvement”.

What is the purpose of commonization? To disseminate best practice.

Which is more important, which takes precedent? Well, I suppose you can supply your own answer to that question, depending on your needs and wishes. It’s clear to me, though, that lean thinking places priority on improvement. Never ending kaizen.

So, my short answer to your question is “no” I don’t see a contradiction. The longer answer is that I see the rationale of the question you raise and the serious concern behind it.

Maybe a good topic for a future Column…

john shook
Frank August 11, 2009
John, yes I would like to encourage you to raise the topic of "standardization in a service environment" in a future column. Reading again your column "Recruiting creative ideas", another thought came to my mind: interestingly the word "standardization" doesn't appear in your column about the suggestion system. Perhaps because standardization happens more on an individual level in manufacturing? Among others, this is highlighted in this section:
"Here’s another of the distinguishing characteristics: the suggestions must (ordinarily) concern YOUR OWN work. No suggestions to improve the cafeteria menu please. We’re looking for ideas to change the process as you are working it."
That might touch a more fundamental difference between manufacturing and services. In manufacturing a specific task is usually performed by relatively few people. There aren't thousands of people mounting the windshield on an assembly line. To simplify: YOUR OWN work is indeed your own work. Hence, dissemination of best practice - commonization - might be less of a challenge.

In service operations, dissemination of best practices is a huge challenge. Think of call center operators giving technical support or decentralized field forces, where it's very hard to deliver a consistent customer experience. We often find thousands of workers performing essentially the same task. The decentral and intangible nature of many services might give a new perspective to standardization. Turning individual high-performance into collective high-performance requires massive commonization. And that in turn forms the baseline for further improvements - the purpose of standardization.

Could it be that standardization and commonization are much closer in service operations than in manufacturing? A second key difference to manufacturing is customer introduced variability in services. See also Frances Frei's paper "Breaking the trade-off between efficiency and service" on http://michaelmyers.biz/materials/HBR-Service-Model.pdf. Would you therefore agree on the hypothesis, that in services it's harder to maintain a stable baseline for improvement, therefore harder to make collective improvements?

Dealing with variability, having a large decentral workforce with homogeneous tasks make improvements, while at the same time maintaining a consistent customer experience is a huge issue in services. That would indeed be a great topic for a future column!
Anonymous August 11, 2009

Your work, and the work of many partners is gaining much traction at Starbucks. I respectfully ask that you do not refer to the management as "kids". This underminds the idea of a "knowledge" worker.

What is especially interesting in the Starbucks case is that they can't even HOPE to do that, not with 10,000 stores in North America managed by kids managing part-timers to provide a high-end product and service.
Lester August 14, 2009
John, Great post. You insight into Lean is always helpful. It is really great that you decided to clarify the story of what is going on at Starbucks, the wild rumors flying in the blogosphere needed to be answered. I am working with the Federal Government on their Lean Transformation now, and am finding it is quite a challenge to interpret the continuous improvement mentality into this setting.... I look forward to your continued posts. Best, Les
Anonymous August 18, 2009
I think Frank hit on a strong point for me which is lack of standardization amongst the organization could possibly be a nightmare. I guess in starbucks case, there is a limited product with massive customization occurring ( I like the comparison to bartending where no 2 bars are alike) What bothers me though is with the assumed high turnover of the baristas, how do you retain best practices? (I remember my 2 weeks with McDonalds…) I have the exact opposite in my industry (automotive parts distribution) with many floor workers with 30+ years in the company. How do you implement standardization and/or commonization with people who have “done things their own way” for so long?
R August 19, 2009
Though lean thinking and actions has proven to be an effective tool for reengineering the mechanics of production processes and systems, it's more cumbersome and difficult to apply (and arguably less effective at improving) the complex human systems that underlie organizational performance.
John August 21, 2009

My main thesis here, or a main thesis, is that whenever you find yourself “chasing best practice dissemination through policing compliance”, you are simply barking up the wrong tree. Yes, I know that presents dilemmas for us. That’s the beauty of lean thinking, really – it helps us work through dilemmas. Or apparent dilemmas. After all, if all this were easy…

Thanks for introducing us to the Frei article. I had read it briefly, but now read it much more closely. I agree with some of Frei's observations and disagree others, including some key ones. I agree with Frei, and you, that service environments present specific challenges to “standardization”. However, I’m not sure at all that, therefore, the distinction I am making between standardization and commonization becomes lessoned. Rather, while yes the interplay between the two probably has a very different dynamic, my sense is that therefore it is all the more necessary to maintain awareness of the distinction.

Specifically to your point about my point from my “Recruiting Great Ideas” column, “…suggestions must (ordinarily) concern YOUR OWN work”, note that I didn’t necessarily mean work that would be done ONLY by the individual making the suggestion. In production, there are often numerous who do the exact same work. Correct that there aren’t thousands of people mounting a windshield, but if there are two shifts and the job rotates every day, there could easily be 10 individuals who do that same job. That always raises issues of agreement among the different people, and compatibility of the suggestion among people with different skill levels, etc.

I will follow your suggestion and write more about this. Thanks for the great thoughts!

John August 21, 2009
Thanks “Anonymous” for sharing your reaction to my use of the word “kids” to describe Starbucks store managers. No disrespect was intended, I assure you. I didn’t think of the term “kid” as derogatory or in conflict with the notion of “knowledge worker” (kids can be knowledgeable, too). Perhaps the other side of 30 looks different when you are late fifties (as I am). I call my son and his colleagues kids even though they are officers in the US Marines. I think the “kids” themselves don’t mind it, but others often mind on their behalf. No lack of respect is intended, but perhaps some is taken, so I shall be more careful.

Mark August 24, 2009
I am 55 retired and work part-time at Starbucks. Lean thinking isn’t just about what is being performed but also about who is doing. At 55 my eyes do not function like a 20 year old and as a result it takes me longer to process a transaction at the register than someone who is half my age. Product labelling (size and location of SKU numbers) needs to be considered when thinking speed. One thing that could really improve the speed of processing orders is to create labels with SKU numbers that someone over 50 can actually read. If the scanner isn’t working (which is often) I need to read the SKU on the package which takes an extremely long time because, for some reason, they are generally printed very small. Also, there is no consistency between products as to where they are placed on the package and therefore; more time is wasted looking for the SKU that I can’t read anyway. Oh, by the way, I love working at Starbucks and affectionately refer to my 25 year old manager as the 'kid'.