注册 登录  
 加关注
   显示下一条  |  关闭
温馨提示!由于新浪微博认证机制调整,您的新浪微博帐号绑定已过期,请重新绑定!立即重新绑定新浪微博》  |  关闭

论老子

道,领导也。领导必需要不断呼唤,教导下属以及以身作则。下属的过和错皆因领导懒惰。

 
 
 

日志

 
 

Chapter 8: Misused of average  

2012-06-24 12:28:12|  分类: Buffer Mentality |  标签: |举报 |字号 订阅

  下载LOFTER 我的照片书  |

On the Friday afternoon, May 23, 2008, I told John, “In order to spot the weaknesses in a VSM report, you need to know how some statistics could be misused or twisted to portray a picture of healthy performance when it is actually not. Statistics is about the processing of data. From a negative perspective, statistics is always being manipulated in favor of the person who compiled it.”

John asked, “What are these statistics?”

I explained, “Well, this is something that I always keep close to my heart. It is not easy to convince anyone how he had misused statistics in his favor. He will defend he has not done anything wrong in the way he had compiled the statistics even though deep in his heart, he had all the intention to manipulate the data to produce a false picture.

Let’s not debate whether this is ethical or not. I prefer to assume this was done in the name of buffer mentality[1]. He who had painted a good picture of his performance had one intention. That is, he wants to buffer himself or his subordinates from the need to work hard to resolve issues immediately.

Let’s start with the statistical average. It is the simplest statistics and most common used. But never in your life did you suspect they can be easily manipulated to buffer against the need to take corrective actions.”

“Wow! This is going to be refreshing. It is going to be a completely different perspective when we look at these statistics. I would like to hear how you review the statistical average,” said John.

 

Illustration #1: Information manipulation at Hewlett-Packard  

 

I explained, “Around mid-June 2005 and in his second month into his new job as CEO of Hewlett-Packard, Mark Hurd split HP's personal computer and printer group back into its two original divisions – it was a merger that was one of the former CEO’s last moves before her departure.

Why would someone merge the two seemingly unrelated divisions together? HP’s former CEO, Carly Fiorina did precisely that.

The printer division is known as the crown jewel of HP. The profit from this division alone was almost equal to the whole of HP’s profit for many years. In other words, the other four divisions of HP added together did not bring in any significant profit. Some of these 4 divisions made money and the rest lost money. Among the loss-making divisions, the personal computer division was the one that lost the most money.

In 2002, Carly Fiorina acquired Compaq – the number one biggest computer maker by sales volume but made zero or little profit for the last several financial quarters before the merger. After the acquisition, the combined personal computer division was showing net losses, quarter after quarter.

By a simple accounting stroke, Carly Fiorina, the incumbent CEO ordered the accounts of the personal computer and printer divisions to be merged as one. This effectively buried the stigma of a continuously bleeding computer division. The net account of these two divisions of course showed profitability all because the printer division is hugely profitable.

The splint in the eye suddenly disappeared. This was purely information manipulation – a behavior that is often used to hide poor performance. The statistical average was effectively being used to erase the poor performance of the computer division.

 

Figure 8-1: Contrasting performance between the personal computer and printer divisions

The same practice was applied to the HP Services division and its IT department. Both of them offer IT services: HP services division provides IT services to clients outside the company but the IT department provides IT services internally to all HP employees. Even though HP Services is a much smaller division in terms of manpower and operating expenses, it brought in substantial revenue. Meanwhile, the IT department was a pure cost center that did not bring in any revenue but cost more than US$2.7 billion a year to maintain.

By combining these two income statements together, the IT department’s annual operating cost of US$2.7 billion did not stick out like a sore thumb[2]. Instead of finding means to reduce its cost of providing IT services to its 151,000 employees, the combined account of the two IT divisions brought in revenue and hence this was used to justify its huge operating expenses.

Around mid-July 2005, Mark Hurd saw through this fallacy and he had hired Dell Inc.'s Chief Information Officer, Randy Mott, to fill the same position he held at Dell Inc.

Further down the line of command, this technique of using the averaging effect of several pieces of data to achieve information manipulation was quite pervasive in HP.

The Tij3.0 printhead consists of two types of pens that go into the commercial printers, namely half-inch and one-inch printheads. Half-inch and one-inch are the respective length of the die that forms the printhead.

In reporting the customer warranty replacement rates for these two printheads, the half-inch printhead performance was consistently below the goal line set at 1%. It was at around 0.8%. The performance of the one inch printhead was consistently above the goal line hovering at around 1.15%.

After combining the warranty replacement rates of these two printheads together, the merged performance was consistently below the goal of 1%. A performance better than the goal means there is no need to take corrective action to reduce the warranty replace rate. This is pure buffer mentality at play.”

“Wow! I did not suspect the statistical average can be used in this manner to hide poor performance. HP employees are rather creative,” said John.

 

Illustration #2: Using historical performance for production planning  

 

“This example is going to be commonly seen in almost all companies. It is the determination of the production capacity used in the production planning process.

Every company has a production planning department. Its primary role is to plan for the quantity of daily output expected out of a piece of production equipment or production cell,” I asked, “How does the production planner obtain the planned output number?”

“From its historical output data, I supposed. Most company will collect the past one or two months’ output data and compute the daily average. With this average number, the production planner aims to load up to this number of quantity for the day’s planned output. I cannot be wrong, can I?” asked John.

“Yes, you are right. This is how all the production planners obtained the average daily planned output number. What does this number mean to you, John?” I asked.

“If I understand correctly, half the time, the actual output is above the average number. Half the time the actual output is below the average number. That is what the average number is all about,” replied John.

I asked, “Half the time the actual output is above the average number and half the time below. Do you get it?”

“Yes, that is right. What is so special about that statement? In the first place, the average number is obtained from the actual output numbers? I don’t get you,” replied John.

I asked, “Let’s take the highest actual output number ever achieved during this say, two-month period. Do you think the factory can repeat the same performance?”

“I am sure it can. It had already proven itself that it could produce that (highest) number of output for one of the days in the two-month period. So it should be able to repeat itself,” replied John.

I added, “I fully agree with you. That highest number of output per day is a proven performance. Certainly the factory can repeat the same performance. But why does the production planner choose the statistical average of all the actual output numbers for this period? And why doesn’t the production planner just pick that highest number of output produced during the two-month period and expect the production folks to repeat its performance at the highest level of output?

I bet the production planner did not suspect any difference between these two numbers. He is merely so used to computing the statistical average. Perhaps, he sees within the organization everybody is using the statistical average to present all their analysis. So he thinks it cannot be wrong to use the statistical average to determine the planned daily output number.”

“Eric, you are such a sharp guy. Never had I questioned why the statistical average is used in production planning and why isn’t the proven highest actual output number used. It is amazing how your mind works,” exclaimed John.

“No, John. To me, everywhere I go, I see buffer mentality. I am so used to finding the presence of buffer mentality in the mind of the employees that where buffer mentality could be, I am sure I can identify it. That has become my instinctive second nature,” I clarified.

John said, “Perhaps, this aptitude makes you stand out form the rest of the lean production grand masters or sinsehs.”

“That is not the end of my knowledge sharing. How do you set the quality goal? How do you set the target for yield losses?” I asked.

John explained, “They are all set with the average number obtained from the recent past historical data. Again, the highest quality level was not used as the goal even though it was already proven it was achievable.

Likewise, the lowest yield loss figure was not used to set the goal for yield loss. The average yield loss number is always the unquestioned number to use.”

I summarized, “Now do you know how rampant the buffer mentality is within an organization. Everyone simply loves the statistical average like it is a God-given number. It cannot be wrong to apply it.

Yet you understand now, how it is being used to give ample room to rest on one’s laurels without pushing himself to work hard at all. This is buffer mentality but in its most inconspicuous form.”



[1] To understand more about buffer mentality, please read “Buffer Mentality” by the same author.

[2] Information manipulation is listed as one of the deadly management behaviors in the book, titled, “7 Deadly Management Behaviors – CEO Edition”.

  评论这张
 
阅读(243)| 评论(0)
推荐

历史上的今天

在LOFTER的更多文章

评论

<#--最新日志,群博日志--> <#--推荐日志--> <#--引用记录--> <#--博主推荐--> <#--随机阅读--> <#--首页推荐--> <#--历史上的今天--> <#--被推荐日志--> <#--上一篇,下一篇--> <#-- 热度 --> <#-- 网易新闻广告 --> <#--右边模块结构--> <#--评论模块结构--> <#--引用模块结构--> <#--博主发起的投票-->
 
 
 
 
 
 
 
 
 
 
 
 
 
 

页脚

网易公司版权所有 ©1997-2017