By Jon Miller | Post Date: November 17, 2009 9:16 PM | Comments: 0
For a group of people who claim to practice management by fact and question the as-given condition, we in the lean community have a troubling habit of citing and accepting made up lean enterprise statistics. In fact I would say that at least 50% of statistics cited about lean have been made up by someone and and passed on to others. I can say this because I personally know some of the people who have made them up.
Here are a few favorites:
In a Manufacturing News interview with former Danaher finance executive and now lean consultant Mark Deluzio, we learn that only 1% of all manufacturers have adopted a lean strategy for growth.
I bow at the feet of the research firm possessing such powers over time and space allowing them to identify and survey "all manufacturers".
When discussing visual management, many have the misconception that any chart that conveys information is a form of visual management. In fact, a made up lean enterprise statistic tells us that 75% of visual controls are just wall paper. This doesn't mean that three quarters of proponents of visual management are laying down wall paper, but that most visual controls are useless. They do not in fact contribute to the immediate understanding of normal versus abnormal, or communicate a standard. This statistic is made up, but grossly true.
A lean company can do the same work as a non-lean company with 50% less space, inventory, equipment and people. Many lean consultants like to shock a prospective client with impressive clients that lean will allow them to get the same work done with half the space, half of the machines, etc. This is correct more often than not, but leaves a lot unsaid, and few who make the claim can substantiate this claim with a sample size of 3,000.
A commonly heard made up lean enterprise statistic during a value stream mapping training or workshop is that less than 5% of lead time is value added. This is a fairly safe statement to make, but given the lack of defined starting and ending points for the lead time, can be simply untrue. It's another shock tactic to raise awareness of how little value added time there is within most traditional end-to-end processes.
Many who hold KPO (Kaizen Promotion Office) positions or equivalent can thank the made up lean enterprise statistic stating that 1% of the workforce should be dedicated to kaizen full time for the job they are in. This is often expressed as 1 full time lean person (KPO) per 100 FTE. This makes planning resources easy, and results in the hiring a lean manager for companies 100 employees or smaller. It is not a very effective way to bring about a culture change. Who knows whether the 1% number is correct or not? If it's correct, why not dedicate 1% of everyone's time to kaizen? In a 100 person company that would be 1% of 200,000 hours per year or 2,000 hours split across 100 people. That's 20 hours per person per year doing kaizen, or 100 minutes per month per person, or 25 minutes per week. By everyone. That. Just. Might. Work.
Some have said that 20% of a group will actively support a change, 20% of a group will actively resist a change, and 60% can be swayed to support or resist. On the one hand this sounds convincing, especially when words such as change management, keys to success and critical success factors are thrown into the sentence around it. However I have heard the same people cite this as 1/3rd, 1/3rd and 1/3rd so the numbers seem arbitrary and rather precise. Also, these numbers skew heavily in one direction or another depending on whether one is in Brazil or in Sweden, depending on what level of leader is leading the change, and whether the group one is addressing has more to gain or more to lose from the change.
The cost of poor quality increases by 1000% each time it is passed downstream. This is sometimes stated as the 10X rule and used to argue for in-process checks, early detection and stop-and-call methods such as andon lamps, jidoka and other aspects of built-in quality. However this made up lean statistic does not hold up when applied indiscriminately. The cost of detecting bad raw material when it is still ore may be thousands of time less than when the finished gear begins to break down within a gearbox within an aircraft. On the other hand an error in a line of software code detected at first quality assurance step may not be significantly higher than if found in the fifth step. The essential thing is that each process is capable of checking both incoming and outgoing quality, and that all defects are caught before escaping to the final customer.
What other made up lean enterprise statistics have you heard of?Comments are moderated to filter spam and inappropriate content. There may be a delay before your comment is published.