Big opportunities that require a big shift in mindset.


As the title plainly says, readers will find below a list of four opportunities for legal industry innovators.  But before we get to the juicy stuff—because everyone loves a list—I want to commend the source for this list: James P. Womack, Daniel T. Jones, & Daniel Roos, The Machine That Changed the World (1990).  It is among my favorite books in any genre and among the most durable contributions to 20th-century management literature.  It is also a fun and fascinating read.

Machine’s story starts in 1950 when a young engineer named Eiji Toyoda visited the Ford Motor Company’s Rouge plant, a facility that turned out 7,000 vehicles a day.  Toyoda’s firm had, in its thirteen-year history, turned out 2,685 vehicles—ever.  Yet after closely studying the most productive factory on earth, Toyoda wrote to his uncle, Kaiichi Toyoda, that he “thought there were some possibilities to improve the production system.”

Today, Toyota is the largest automaker in the world and earns four times as much profit per vehicle as any major American or European brand.  This is because the Toyota product commands a price premium even though it is manufactured at a lower unit cost.  What is the reason for this stunning market domination? Process supremacy.

The Machine That Changed the World unpacks that story.  It is a catalog of the vexing inefficiencies and process problems that plagued a mature and highly profitable industry.  It is a review of the solutions to those problems that collectively came to be known as the “lean production method.”  And, what may resonate most with legal innovators, it is a terrifying account of how Detroit resisted and ultimately rejected nearly every successful innovation coming from Japan.  (For an illuminating and at times jaw-dropping account of the tragic organ rejection Toyota’s ideas experienced when transplanted into the American auto industry, spend an hour with This American Life, Episode 561: NUMMI. See also Post 057 (Henderson discussing use of NUMMI case study in law school leadership class.))

Machine is thus a book-length indictment alleging that if you ever had an American car that sucked, it wasn’t because Detroit didn’t know better.  It was because Detroit refused to do better.

The same indictment could be filed against the legal industry in 2020, see Post 159 (Jason Barnwell comparing the law to a “chaotic factory” for knowledge work), and that makes Machine powerful reading for legal industry innovators.  Our problems can be identified, and the basic solutions are known from lessons learned in other industries.  The remaining question is how the industry will carry those solutions into practice.  Will incumbent firms fail like Chrysler—though without the benefit of a bailout?  Will some incumbents reorder themselves to become the Toyota of law?  Or will disruptors transform existing structures and introduce something entirely different?

There are of course a few legal industry organizations that truly “get it” – e.g., my former employers at Novus Law, who “get it” the most and for whom nothing I have written here will be news.  But the vast majority of law firms and other legal industry players will continue to discover problems just as Detroit discovered them, and they will fight just as hard as Detroit did to reject the innovations necessary to solve them.

With that, I offer investors, innovators, and incumbents some grist from the mill supplemented with page references to the hardcover edition of The Machine That Changed the World.

Opportunity #1—Expose and reduce Law’s hidden error rate (Machine, pp. 53-57)

I believe the story of Toyota begins with the following insight: The moment an error is made, it is an error forever.  It’s cost and quality implications can never be erased, even if it is “fixed”.  Viewed this way, the typical Ford (or Mercedes or Fiat) vehicle experienced hundreds of errors in the course of its manufacture.  Some of these errors would later be identified and fixed—in the re-work area of the plant, or by the dealer, or by the unlucky owner.  This wave of errors was caused by a belief that all work on a vehicle was productive work, whether it was manufacturing a car for the first time or performing re-work after it came off the line.

By contrast, Toyota treated every unnecessary exertion of energy as an error (muda), because even if caught and fixed, an error reduces efficiency and therefore profitability.  In truth, most errors in a complex product are never found and fixed, so the ultimate cost of most errors is a defective product.  With this very different attitude to what constitutes quality and how to define an error, Toyota learned to treat every error as a “stop work” moment—an opportunity to root cause why the error occurred and to eliminate the cause.  This is an annoying disruption to workflow in the short-run, but it radically increases quality in the long-run.

Seen through this lens, the law has a grave quality problem that shows up in customer cost, law firm profitability (acutely in the AmLaw 200), and substantive legal outcomes.  I have long thought that the beginning of true legal industry reform will be the widespread acceptance that even an elite law firm’s best output is riddled with errors because errors are properly defined as any wasted effort.  We are nowhere near the acceptance of this fundamental point because, like Detroit in 1980, we have simply defined quality to exclude the profusion of obvious errors lawyers commit every day.  This collective blindness is perfectly normal, an act of motivated reasoning undergirded by the security and success of the industry’s leading lights.

To be clear on how radical the lean production insight is, let me enrage at least a few readers by making the following claim: every edit to a legal document is muda—it is wasted effort, and therefore an error to be prevented—because every edit flows from the failure to get it right the first time.  Yes, there are differences between an allonge and a RAV4, but from a process perspective, they are both complex products with knowable supply chains and definable quality characteristics.  It may be hard to get each step right the first time, but nobody can deny that it is possible.  Law will begin its true reform in earnest when the industry accepts (or, more likely, is forced to accept) the daunting task of trying to get everything right the first time.  

GM thought that the task was folly.  Toyota didn’t.

A mundane illustration of the law’s hidden errors in practice: I first encountered muda in earnest at about 2:00 a.m. while working with two partners on a summary judgment brief in the Spring of 1998.  The red-marker-on-paper edits of one partner rejected what had, just an hour before, been the red-marker-on-paper edits of another partner.  This was, if memory serves, in about the 18th footnote.  At that point, I realized that the team had fought itself to a standstill, and yet we greeted the rising sun with the hum of printers in the background.  Dueling edits carried on through breakfast.

While these edits may eventually have resulted in a slightly better brief than the first draft, (a) we had no mechanism for knowing that, and (b) it was obvious that most of the energy expended was not economically productive in total, in the sense of delivering marginal gains in excess of marginal costs.  Most importantly, (c) the right response to this experience is to design in advance an assembly sequence that pre-emptively eliminates re-work.  

This idea is so alien to law firm culture that any expression of the point will meet with blank stares.  It was like that at Chrysler, too.

The innovator’s opportunity here is to realize that, while law has a strong cultural resistance to identifying and fixing its illimitable hidden errors, tremendous profits await those who expose and reduce the hidden error rate.  AmLaw 200 firms are now forced to write off most of their wasted effort, and even AmLaw 20 firms could earn more by eliminating muda and realizing a price premium for increased product quality and reliability.  As Machine shows on nearly every one of its 278 pages, the failure to put process in the foreground is the root cause of wasted effort, and wasted effort is always and everywhere the source of cost and quality problems. 

Opportunity #2—Create discipline through error feedback loops (Machine, p. 62)

Detroit carried tons of inventory of every part at all times, meaning that a defective part was really no problem—skip over it and grab a good part from the bin.  This allowed defects to grow like weeds. Of course, this resulted in many defective parts being embedded deep into new vehicles.

Toyota ran towards this problem rather than away from it: in its kanban supply system, a bin of parts was delivered to each work-station, and it contained only (and exactly) enough parts to finish the job at hand.  If a single part was defective, this was a production-stopping moment, forcing accountability to immediately flow back upstream to the source of the defective part.  This was done on purpose and for this reason.  The accountability of this immediate feedback led to a production system nearly devoid of errors, even errors defined broadly to include wasted effort.

This kind of immediate feedback is painful—and effective.  Across many years of hockey, my best and most effective coach had a whistle in his mouth throughout every practice.  He was ruthless to shut down every drill the moment an error was made; every player froze in place as we diagnosed the error.  We did it again until it was right.

Toyota placed scarcity-by-design at key points in its production system. Those scarcities forced an immediate shutdown in response to an error, intensifying its immediate cost and making it impractical not to root out the error’s cause right away.

Law has no scarcities of this kind, no discipline of immediate stoppage.  Errors inside a process are met with no immediate consequence, so they are repeated an unknown number of times with only informal methods of resolution.  Efficiency and profitability gains await the innovators who design a process that demands its own perpetual reform.  If it followed the Toyota model, this practice would interrupt the whole process when an error occurs so the root cause can be identified and eliminated.

This feedback loop is really what made Toyota great—the perpetual necessity of identifying and fixing errors.  The increasing precision and decreasing tolerance for error would eventually result in machining so precise that in some cases Japanese parts wouldn’t even fit “compatible” American components, because Detroit’s margin of error was necessarily so forgiving and Nagoya’s margin of error had essentially disappeared.  (For a discussion of the “five whys” approach to root causing process error, see Machine at 79.)

God willing, someday a legal organization will have the problem of too much precision that results from a commitment to never making an error the first time around.

Opportunity #3—Empower your players to reform the process (Machine, pp. 42, 53 and 92)

It is unproductive to identify the root cause of errors if the team is not empowered to fix them.  Machine discusses Detroit’s problem here: workers were regarded as little more than interchangeable parts (p. 42).  Since Toyota’s system stopped at the appearance of every material error, it couldn’t wait long for a solution.  The minds and hands that made the cars were necessarily the same ones entrusted to fix the process.

Lawyers, paralegals, technologists, and other key players are rarely engaged in designing processes in most legal industry firms.  There are lots of reasons for this, including that all of this activity would be non-billable in the conventional law firm business model.  But there’s an opportunity for the firm that creates superior outcomes (as defined by both cost and quality) by empowering key players to fix the problems they see when they see them.

A naturally related point is that any firm committed to empowering its production staff will have to commit to their training and development at a level that goes deeper than just rote learning.  The Japanese and American approaches were in stark contrast as of 1990: 380 hours of training per employee, per year in Japan; 46 in America (Machine, p. 92).  Admitting that this is a very rough input proxy, if we ignore CLE (ahem), how many hours of value-adding training do incumbent legal industry firms typically provide?

Opportunity #4—Increase the “supplier share of engineering” (Machine, pp. 59-60, 118, 167)

Before the lean production revolution, automakers had been allergic to components not designed within their own walls.  Suppliers could manufacture but were not trusted with engineering.  This isn’t just a car industry problem—every organization has strong antibodies to something “not invented here.”

Yet many accountability and efficiency problems flow from this kind of tribal behavior.  One can only imagine the cumbersome communication loops that took place when a defect was discovered on the production line in Detroit and eventually determined to have originated with a third-tier supplier in Ontario due to an engineering miscalculation that took place in Fremont.  How many (thousands of) defective cars rolled down the line while that problem was rectified, assuming it ever was?

Again, contrast lean production methods: 

[Suppliers] were given a performance specification.  For example, they were told to design a set of brakes that could stop a 2,200-pound car from 60 miles per hour in 200 feet ten times in succession without fading.  The brakes should fit into a space 6” x 8” x 10” at the end of each axle and be delivered to the assembly plant for $40 a set . . . Toyota did not specify what the brakes were made of or how they were to work.  These were engineering decisions for the supplier to make.  (Machine, p. 60)

Suppliers are immediately held accountable for product quality as if they are on the shop floor – but that accountability still respects their design authority.  In the mid-1980s, when Nagoya’s quality advantage over Detroit was at its height, Japanese suppliers had a 51% share of engineering; American suppliers had 14%. (p. 118) If this sounds inefficient (as Detroit thought it was), consider that a Japanese car at the time had a total design team including suppliers of 485 engineers, as opposed to 900 in America, and “a totally new Japanese car required 1.7 million hours of engineering effort on average and took forty-six months from first design to customer deliveries” while a comparable American vehicle took 3 million hours and took sixty months.  (p. 111)  

Did I mention that Toyota was more profitable due to its efficiency?  This kind of efficiency only existed because there was no duplication of work between suppliers and Toyota.  That required that Toyota trust its suppliers to do real value-added work, not just to stamp out parts.

As law matures into a supply chain, consider how little a senior lawyer genuinely trusts his or her suppliers.  How much true ownership resides with an e-discovery, analytics, technology, or contract lawyer supplier?

In its most primitive form, the extended legal enterprise consisted of contract lawyers for rent and e-discovery firms that followed orders.  Their work was always reviewed closely and not infrequently entirely re-done by law firm associates, all at the client’s cost, and with no mechanism in place to demonstrate that the re-work would result in higher quality.  The quality control problems, if quality control was even imagined to be a discipline in this context, were legion.

I was generous to use the past tense in the preceding paragraph.  It’s mostly still like this in 2020.  We have seen no serious movement on this point in 30 years.

There is an opportunity to create efficiencies at the heart of legal matter execution for any firm that vests true ownership in its supply chain—in plain English, by delegating outcomes rather than delegating tasks.  As phrased above, we need to increase the supplier share of engineering.  This requires two things: clear outcome specifications (“we need brakes of these precise dimensions with these performance characteristics”), and the cultural ability to trust suppliers.  

On the latter point, senior lawyers will usually reject this approach with three words – “it’s my license”.  

This is exactly wrong.  “Their license” is overwhelmingly about error prevention, so what serves “their license” is the path that reduces error through the best available means.  Suppliers that are trusted, empowered, and held accountable are far more likely to reduce variability.  This is a commander’s intent approach that puts quality and effectiveness directly on the table and holds the whole supply chain accountable for the final result.

A Final Word

I want to conclude by addressing an objection that arises when one uses manufacturing examples to elucidate process problems in the information economy.

As discussed above in the section on supplier engineering, the efficiencies captured by Toyota’s process discipline were not only—and not necessarily mostly—on the manufacturing floor.  Toyota’s vehicles were not just assembled far more efficiently; they were designed far more efficiently too. Car design blends art and science just as law does, and it is entirely complex and collaborative information work. The process disciplines necessary to efficiently design and engineer the Corolla would be much the same as the process disciplines (theoretically) necessary to efficiently divest the subsidiaries of a multinational company or to construct a summary judgment brief in a complex case.  If anything, the Corolla project involves more subjectivity and complexity and has more potential for error and redundancy than those two legal examples.

Lawyers will protest that they are necessarily craftsmen, that their judgment is at the heart of the work.  And they are right, but that is even more true of car design.  When Toyota gets better design out of a third less work than GM, Toyota wins and GM goes bankrupt.

The examples in Machine tend to come from the manufacturing floor because process is most visible there and errors are easiest to show. But the data collected by the book’s research team related as much to the information economy inside the auto industry as to the physical economy of the plant. The book’s lessons are as powerful for those working with keyboards as they are for those working with robots and wrenches.