GM’s Valukas Report | Lean in the U.S.A.

General Motors has long had a reputation as the antithesis of Toyota, so GM news is sometimes of interest to Lean practitioners. The recently released Valukas report on GM’s ignition switch problems is a case in point.

The GM board brought in a Chicago law firm to perform an independent investigation of how the ignition switch problem developed and to document the actions and reactions, if any, of corporate officers and management. The board further charged the investigators with making recommendations to preclude similar failures in the future. The investigators were lead by former U.S. attorney Anton Valukas, hence its moniker “the Valukas report”.

In addition to its interest to Lean practitioners, the report should also be of interest to anybody who periodically deals with an intransigent private sector bureaucracy and is driven to bang his or her head against the wall out of frustration. For managers in both government and business the report illustrates in horrifying detail how far things can go wrong when we let them. For Lean practitioners it provides a base point for a reality check; reflective people ought to ask themselves if they are confident GM’s problem couldn’t have happened to them.

I highly recommend reading the report. (Be warned that the .pdf file is about 325 pages, so it’s a long download. It is, however, double spaced with many footnotes, so it’s a shorter read than it first appears.)

******************

Five big reasons you need standard work

Mass production operations often have what they call work standards, which are really nothing more than quotas. In many cases, whatever you do to fill that quota is fine, which is why mass production operations can put out such poor quality and be such barbaric places to work.
Lean operations have what sounds like a similar sort of thing: standard work. However, standard work and work standards are very differnt.
The work standards of mass producers are narrowly outcome focused. If upper management dictates that a barrel must be filled with bolts by the end of the shift, that barrel will be filled with bolts, or something roughly resembling bolts, by the end of the shift no matter what it takes.
In contrast, standard work is process focused. When we design and use the right process, there will be no problem producing enough well-made bolts to fill the barrel, if a barrel is even the right way to measure, pack, and/or ship bolts. The right process will produce the right result. Mass producers skip designing and using the process. They fixate on the result, leaving them little ability to actually influence the result. That’s why they so often have long lead times, poor reliability, poor quality, and poor relationships between management and line level employees.
That said, there are five big reasons you must have, and use, standard work in a Lean organization.
1) employees become resistant to change
2) Deming said so
3) statistical control impossible without it
4) Inputs into next process become random
5) it helps identify problem personalities.
Big Reason #1: If you don’t teach new employees the best way to perform a task, they will figure out how to do it on their own. This is just common sense.
The problem is that these new employees, or employees that have transferred from another part of the organization, won’t figure out a way of performing the task that uses the fewest resources. They won’t figure a way that takes the least amount of time. They won’t figure out the way that produces the highest quality. Instead, they will figure out the way that is easiest for them to figure out.
Now, having figured out the way that is easiest to figure out, they will instantly become resistant to change.
In the long run, in a business that does this, virtually every processes in the business will have been designed by people who, through no fault of their own, knew nothing and cared little about the the needs of the customer, the needs of the business, the resources available, or the system generally. This is a recipe for seriously suboptimal performance.
The best way — really the only way — to avoid this problem of suboptimal processes wedded to a resistance to change is to have standard work defined for every task, to train employees in standard work from the very beginning of their employment, and then expect them to use it. Of course, it’s also necessary that they learn that suggestions for better ways to do the job are welcome, and it’s very likely that the standard work will change in the future based on those suggestions, but for now they must master the existing standard work.

Big Reason #2: Edwards Deming said so.
Recall that there are four parts to the Deming cycle that Deming described in his famous 1950 speech in Hakone, Japan: customer survey, design, production, and sales or delivery.
Insert graphic here
The design step in this cycle isn’t merely design of a product or service. Design includes design of the processes used to produce the good or service.
If you doubt that design of the process is important, consider any number of tech companies: Amazon.com, Ebay, and Google, for example. None of these companies actually produce physical products. The only thing they can design is process. Design of a process means, in large part, defining the standard methods by which value is created. Standard work, then, is an integral part of the Deming cycle which essentially defines Lean. No standard work, no Lean.
Big Reason #3: Statistical control is impossible without standard work.
Lean process improvement is often heavily dependent on the use of statistics to measure the state of a process. If you don’t have and use standard work, if every employee is doing their own thing, you don’t have a process to measure. Instead, you have as many processes to measure as their are employees doing their own thing.
If you treat this mass of processes as one process, it’s extremely unlikely you will get good results from any statistical analysis. If you try to make decisions based on such faulty analysis, you will simply compound the problem of not having standard work.
Big Reason #4: It’s the GIGO principle: Garbage In, Garbage Out.
Remember that any organization is a system — a collection of components that interact to produce an outcome. A large portion of the interactions in a system are customer/supplier relationships. When you lack standard work in any one component, that component produces random outputs, which become random inputs for the next component in the system. If this second component of the system is able to correct the random problems supplied by the first, it suggests that the first component could have solved the problem itself. It also suggests that

Each component Random inputs lead to random outputs.

into next process become random