The Southwest Airlines Way

Southwest_Airlines_WayCHAPTER11.pdf

Home >Business & Finance homework help >Operations Management homework help >The Southwest Airlines Way

C H A P T E R

139

Avoid Finger Pointing— Measure Performance Broadly

You can get to the point where you saturate yourselves with information, and you get paralyzed. We have more interest in broader categories, rather than analyzing and assessing blame. It’s easier to adjust with broader categories. Maintenance could come up with 50 categories of flight delays, if they wanted to. But you end up chasing your tail.

—Jim Wimberly, Executive Vice President of Operations, Southwest Airlines

LI K E M A N Y O R G A N I Z A T I O N S in other industries, airlines have traditionally relied on systems of functional accountability. Out- comes of the departure process are typically divided into departmental objectives, for which individual departments are held accountable. Each departure delay is traced to the department that is thought to have caused it. Then, on a daily, weekly, and monthly basis, the percentage of on-time departures is calculated for each department.

This system of accountability tends to generate a search for depart- mental failure. Yet because of the task interdependencies in the flight departure process, it is often difficult to determine which department caused a particular delay. One rule of thumb often used is “whoever was off the plane last.” If the gate agent who was boarding passengers was last

Chap 11 11/20/02 3:03 PM Page 139

Gittell, Jody Hoffer. Southwest Airlines Way, McGraw-Hill Trade, 2000. ProQuest Ebook Central, http://ebookcentral.proquest.com/lib/erau/detail.action?docID=4654913. Created from erau on 2022-09-17 04:20:58.

C op

yr ig

ht ©

2 00

0. M

cG ra

w -H

ill T

ra de

. A ll

rig ht

s re

se rv

ed .

off the plane, it is presumed to be a customer service delay. If the ramp agents loading baggage were last off the plane, it is presumed to be a ramp delay. If the fueler was the last one off, it is presumed to be a fuel- ing delay. Therefore the common pattern is a race to finish one’s own assigned task before the other groups finish their tasks, even when coop- eration between the groups would improve the speed and quality of the process. Worse, participants tend to hide information to avoid blame, thus undermining the potential for learning.

Through these unintended dynamics, functional accountability un- dermines relationships of shared goals, shared knowledge, and mutual respect among those who must coordinate their tasks in order to achieve high performance. To achieve quality outcomes in the face of weak coor- dination requires longer turnaround times and higher staffing levels, resulting in tremendous efficiency losses. This chapter shows that there are constructive alternatives to these systems of functional accountability. Cross-functional performance measures encourage participants to focus on learning rather than blaming when things go wrong, thereby bolstering relationships of shared goals, shared knowledge, and mutual respect. We will see how Southwest Airlines in particular has learned these lessons.1

Performance Measurement at Southwest Airlines

Determining the cause of a delay had once been a conflict-ridden process at Southwest, as it was at other airlines, and had often deteriorated from problem solving to finger pointing and blame avoidance. Southwest countered this tendency in the early 1990s by instituting a “team delay” which allowed less precise reporting of the cause of delays, with the goal of diffusing blame and encouraging learning. According to Jim Wim- berly, executive vice president of operations:

We had too many angry disagreements between flight attendants and gate agents about whose delay it was. It was too hard to determine whose fault it was.

One of Southwest’s chief pilots explained:

The team delay is used to point out problems between two or three different employee groups in working together. We used to do it [in the following way]: if people were still in the jet way at departure time, it was a station delay. If people were on board at departure time, it was a flight crew delay.

140 T E N S O U T H W E S T P R A C T I C E S

Chap 11 11/20/02 3:03 PM Page 140

C op

yr ig

ht ©

2 00

0. M

cG ra

w -H

ill T

ra de

. A ll

rig ht

s re

se rv

ed .

But now if you see everybody working as a team, and it’s a team problem, you call it a team delay. It’s been a very positive thing.

In addition to the team delay, Southwest has about 10 other delay cat- egories, far fewer than at other major airlines. The reduced precision of performance measurement did not appear to concern Southwest leaders. How does Southwest motivate performance? A station manager explained:

Through personal pride. Because we’ve always done it, I guess. Also, we track the source of delays. Usually it’s a situation rather than a person who is at fault. We take a delay when the situation warrants it. Besides this, we have delay codes to identify which department caused the delay. We try to figure out what caused a delay, but we don’t do much finger pointing. We find that the more you point fingers, the more problems go underground rather than getting solved.

The only punitive measures are taken when there’s a personnel delay. When someone just wasn’t there to do their job. Supervisors often refer disci- plinary problems upwards to the department manager, if they can’t solve them.

Southwest also uses rewards to motivate performance, he explained:

Customers send letters to headquarters, with compliments or complaints, about 5000 per month. These letters are sent to the relevant station, then when I get it I will put a smiley face sticker on it and frame it. People like to see their name up there.

We have agent of the month awards in each department. The winners are chosen by their fellow employees. Then managers and supervisors pick agents of the quarter from among the agents of the month. All agents of the quarter come to an award lunch to receive their plaques.

We also use $5 meal vouchers to reward people for good performance. Supervisors do this. And agents reward each other by sending Love Reports.

Benefits of Cross-Functional Performance Measurement

Relative to the other airlines included in this study (United, Continental, and American), Southwest took a cross-functional approach to perfor- mance measurement.2 This study, outlined in detail in Chap. 3, included two American Airlines sites, two Continental Airlines sites, two South- west Airlines sites, and three United Airlines sites. A statistical analysis

Avoid F inger Point ing—Measure Per formance Broadly 141

Chap 11 11/20/02 3:03 PM Page 141

C op

yr ig

ht ©

2 00

0. M

cG ra

w -H

ill T

ra de

. A ll

rig ht

s re

se rv

ed .

showed that a cross-functional approach to performance measurement is associated with higher levels of relational coordination. Cross-functional performance measurement also contributes to improved flight departure performance, particularly faster turnaround times, greater staffing pro- ductivity, fewer lost bags, and fewer customer complaints.3 To observe the effects of performance measurement on relational coordination, we can plot performance measurement for each of the nine sites against relational coordination. Exhibit 11–1 suggests a clear impact of cross- functional performance measurement on relational coordination.

Performance Measurement at United Airlines

Up through the mid-1990s, United had a system of functional account- ability for delays, similar to the rest of the industry. As a result, a Boston ramp manager explained:

142 T E N S O U T H W E S T P R A C T I C E S

* Each circle denotes one of the nine sites included in the study. Cross-functional perfor- mance measurement is measured as the number of functions that could be held jointly accountable for a delay. Relational coordination—coordination carried out through rela- tionships of shared goals, shared knowledge, and mutual respect—is measured as the per- centage of cross-functional ties that are strong or very strong, based on an employee survey. Relational coordination in turn has a positive impact on quality and efficiency per- formance, as illustrated in Exhibit 3–4.

Relational Coordination

C ro

ss -F

un ct

io na

l P er

fo rm

an ce

M ea

su re

m en

t 7

28% 62%

CON2 CON1

UAL3

UAL2 UAL1

SWA1SWA2

AMR1AMR2

Exhibit 11–1 Impact of Performance Measurement on Relational Coordination*

Chap 11 11/20/02 3:03 PM Page 142

C op

yr ig

ht ©

2 00

0. M

cG ra

w -H

ill T

ra de

. A ll

rig ht

s re

se rv

ed .

At United—and I’m sure at others—there was always a lot of finger pointing at different departments and different divisions. All divisions had their own goals. They weren’t interconnected. The attitude was, if they are taking a delay at least it’s not mine, so you would sort of forget about it.

Under this system, assigning delays involved a great deal of conflict, according to the Los Angeles operations manager:

You used to need a titanium suit. We used to spend hours and hours figuring out whose delay it was.

Reinforcing these dynamics within each station was a Friday conference call with the regional director, nicknamed the “Friday Flogging.” Ac- cording to the Boston ramp manager:

Our regional vice president is from the old school of management. You hate to manage by intimidation. Our vice president is more along these lines than others. We’re changing over.

On the theory that “if you’re responsible for all delays, you might get more involved in the process and help them out,” United introduced a new delay accounting system in the mid-1990s. According to a Los Angeles customer service manager:

Individual managers are not responsible for just their own department’s delays because we have families of delays now. As a customer service man- ager, I may be responsible for delays that are partly caused by flight atten- dants. This means I’m supposed to communicate with that other group. Flight attendants and customer service agents interact a lot. This system makes them talk. It’s a family of delays. There is no win or lose.

Under this new system of accountability, up to three departments can be held responsible for a given delay, and all are therefore expected to work on finding and resolving the underlying problems.

The new system also includes a focus on on-time arrivals rather than departures, in effect combining the goals of the station and the flight department. In keeping with this, a team delay code was added to allow the station and flight department to take joint responsibility for a delay. The team delay was to be used exclusively, however, for getting additional

Avoid F inger Point ing—Measure Per formance Broadly 143

Chap 11 11/20/02 3:03 PM Page 143

C op

yr ig

ht ©

2 00

0. M

cG ra

w -H

ill T

ra de

. A ll

rig ht

s re

se rv

ed .

revenue on the flight—for example, additional passengers or freight. It is not supposed to be used to take joint responsibility for breakdowns in communication, as it was at Southwest.

The Shuttle took this team delay concept a bit further than the rest of United Airlines, however, in keeping with its focus on teamwork. Whereas three functions could be held jointly responsible for a delay in mainline United, up to four functions could take responsibility in the Shuttle operations.

Due in part, perhaps, to the Shuttle’s approach to performance mea- surement (and other practices highlighted in this book), the United Shut- tle achieved significantly higher levels of relational coordination compared to United’s non-Shuttle sites (55 percent versus 42 percent of cross-func- tional ties were rated by employees as strong or very strong). Exhibit 11–1 shows how these higher levels of relational coordination are related to the Shuttle’s performance measurement practices. The United Shuttle, denoted by UAL3 in Exhibit 11–1, has higher levels of relational coordi- nation than its non-Shuttle counterparts (UAL1 and UAL2), correspond- ing to its more cross-functional approach to performance measurement. In addition, as we saw in Chap. 7, the Shuttle outperformed its non-Shuttle counterparts at United on both quality and efficiency dimensions.

Performance Measurement at Continental Airlines

Continental Airlines remained more specific in its delay reporting, allowing each delay to be charged to only two departments. An opera- tions coordinator explained:

We can split the delay into primary and secondary causes. We split the min- utes and charge them to different departments.

The possibility at Continental of charging two departments rather than one raised the level of shared responsibility somewhat, but still the measurement system was perceived to be an obstacle to coordination and teamwork. According to the Cleveland station manager:

We might take one delay and take one or two hours to find the root cause of the problem. Instead of punitive action, we use a positive constructive approach.

144 T E N S O U T H W E S T P R A C T I C E S

Chap 11 11/20/02 3:03 PM Page 144

C op

yr ig

ht ©

2 00

0. M

cG ra

w -H

ill T

ra de

. A ll

rig ht

s re

se rv

ed .

However, interviews with supervisors and frontline employees re- vealed that this “positive constructive” approach was counterproductive in certain respects. It encouraged them to pay excessive attention to doc- umenting who did what at what time, rather than looking ahead to figure out together what should be done to get the next plane out.

Also, despite the efforts to focus on objective reporting of what hap- pened, fear reportedly continued to play a major role. According to an operations coordinator:

There’s this fear over taking a delay—everybody fears they’ll be chastised for it. You spend so much time filling out delay forms and fighting over a delay—just think what we could be doing. We had a two-minute delay that no one would take responsibility for.

According to the station manager:

Barriers between groups—it all comes down to the delay coding system. Upper management just wants to have a tracking system. If you have a lot of code 31s (maintenance delays), and then the maintenance guy is gone, you know it is punitive. That is the bottom line.

You come in front of a tribunal. Headquarters doesn’t have time to look at the details. They just see code 10s and then passenger services has a lot of explaining to do. . . . It’s a punitive system, but if you’re a good station man- ager, you buffer it, don’t allow it to cascade down to the frontline employees.

Comments from frontline employees, however, suggested that this at- tempt to buffer them from headquarters’ punitive approach to account- ability had not fully succeeded.

Exhibit 11–1 shows Continental’s functional approach to perfor- mance measurement, relative to Southwest Airlines, and its correspond- ingly lower levels of relational coordination. We see little variation between the two Continental sites in their approaches to performance measurement. Headquarters often resists local variations in performance measurement, as we will see in the case of American Airlines, leaving local managers to buffer employees as much as possible from their potential negative effects.

Avoid F inger Point ing—Measure Per formance Broadly 145

Chap 11 11/20/02 3:03 PM Page 145

C op

yr ig

ht ©

2 00

0. M

cG ra

w -H

ill T

ra de

. A ll

rig ht

s re

se rv

ed .

Performance Measurement at American Airlines

American Airlines exemplifies the attempt to achieve accountability through a strictly functional approach to performance measurement. One field manager explained the company’s philosophy, as he under- stood it:

It helps a lot just to keep score. People are naturally competitive. They absolutely need to know the score. Once they know, they will do something about it. Every delay comes to my attention and gets a full investigation. . . . The last thing most of them want is the spotlight on them. I just increased the amount they had to do to keep the spotlight off of themselves.

Each time a delay occurred, managers on duty were responsible for figuring out which function caused it. Immediate penalties accompanied delays, in the form of having to explain what happened. If a delay occurred on a flight scheduled to make connections elsewhere, “Crandall wants to see the corpse,” said a ramp manager in Boston. “It is manage- ment by intimidation.”

This system had the unintended effect of encouraging employees to look out for themselves and avoid recrimination, rather than focusing on their shared goals of achieving high-quality outcomes efficiently. A ramp supervisor explained how goal displacement occurred:

If you ask anyone here, what’s the last thing you think of when there’s a problem, I bet your bottom dollar it’s the customer. And these are guys who bust their butts every day. But they’re thinking, how do I keep my ass out of the sling.

American’s system of accountability also resulted in a great deal of time spent trying to sort out the cause of delays. According to an opera- tions manager:

There is so much internal debate and reports and meetings. This is time that we could be focusing on the passengers.

Another result was frequent misidentification of the problem. A ramp supervisor told me, only half in jest:

146 T E N S O U T H W E S T P R A C T I C E S

Chap 11 11/20/02 3:03 PM Page 146

C op

yr ig

ht ©

2 00

0. M

cG ra

w -H

ill T

ra de

. A ll

rig ht

s re

se rv

ed .

We have delay codes for when the Pope visits, or if there are beetles in the cockpit, but sometimes a problem occurs routinely and we have no code for it. What usually happens is a communication breakdown, but we have no code for that. So we tag it on the last group off the plane.

Sometimes there is a failure to identify the problem altogether, due to outright distortion of information. Due to the perceived harshness of consequences for delays, cheating sometimes occurs—for example, the practice of releasing the aircraft brake early to prevent a delay from being registered electronically. According to a pilot:

Gate agents are so much under the gun. They are scared of disciplinary action. Some of us will tell the gate agent—take all the time you need to load this plane. I’ll take care of it. Captains have authority. We can use it well. If you use your authority as a captain to take the fear away from the gate agent, they can relax and do their job. It’s informal, behind the scenes. We were dis- torting the data to keep the gate agent from getting in trouble.

In this example, employees collaborated across functional boundaries to mitigate the perceived harshness of the performance measurement system. Most often, however, this system of functional accountability undermined relationships across functional boundaries. A ramp supervi- sor in Los Angeles complained:

It seems like it should be very easy to sit down and develop a plan for every- one to work as a family. It’s because of all the report cards, the cover your ass stats, the cover your boss. You’ll do whatever you can. . . . How can everyone work together, striving to be perfect, when it’s going to be on your report card, reflect on your performance?

Station managers at American were intimately aware of these prob- lems, and attempted to solve them through changes in the system of per- formance measurement. According to American’s station manager at Boston, rather than hold people individually accountable for their con- tribution to outcomes, the company should hold people collectively accountable for outcomes.

Avoid F inger Point ing—Measure Per formance Broadly 147

Chap 11 11/20/02 3:03 PM Page 147

C op

yr ig

ht ©

2 00

0. M

cG ra

w -H

ill T

ra de

. A ll

rig ht

s re

se rv

ed .

In a company that depends so heavily on cross-functional teamwork, how useful is a diagnostic system that focuses on individual efforts exclusively and does not effectively measure the ability of the team to work together?

She eventually tried to implement cross-functional teams and cross-func- tional measures of performance at her station. She received no support from the company for this initiative, however, and quit in discouragement after 14 years of service to the company.

Soon afterwards, ironically, American Airlines launched a different initiative to move toward a more cross-functional approach to perfor- mance measurement. Departure teams were formed wherein pilots could determine, in conjunction with gate agents and other employees, whether to delay a flight up to 5 minutes to accommodate passengers or their baggage. This was called a “discretionary delay” program. Some pilots agreed they had much insight to offer based on their trips from sta- tion to station. “The captain sees all,” one pilot told me, “and can make a lot of recommendations for making things work better.”

However, pilots were limited in their ability to work as part of a team with other employee groups. Rather than deciding in conjunction with the other employee groups, pilots reportedly insisted on being the first among equals. According to the vice president of flight operations:

There are real problems with the way that program is working right now. The pilot thinks he is in total control and that the ground workers don’t know as much. The gate agents are getting around the pilots by cheating, saying they already got approval from the pilots when they didn’t.

The senior vice president of operations concurred, claiming:

The performance is dismal; dismal. We are doing the customers a disservice. The lesson is that captains without the right knowledge base cannot make decisions properly.

Pilots were limited in the decisions they could make, according to these managers, because they did not understand the perspectives of the other groups well enough to make the relevant trade-offs. The initiative was considered a failure and dropped.

Other local innovations were also tried, with little success. Consistent with the thinking of the Boston station manager, the Los Angeles station

148 T E N S O U T H W E S T P R A C T I C E S

Chap 11 11/20/02 3:03 PM Page 148

C op

yr ig

ht ©

2 00

0. M

cG ra

w -H

ill T

ra de

. A ll

rig ht

s re

se rv

ed .

manager initiated a system of cross-functional accountability for delays, in which up to three functions could take joint responsibility for a delay. He explained:

Delays are usually caused by a combination of factors—say a lack of commu- nication between the gate and cabin cleaning. So we came up with a system for coding delays jointly to the groups that contributed.

However, the experiment fell apart when American’s headquarters insisted the station would have to continue identifying and reporting to headquarters the one functional group that was responsible for the delay. According to this station manager:

It undermined the credibility of our program here. People didn’t believe anything had changed.

Like the Boston station manager, he left the company soon after. Exhibit 11–1 shows American’s highly functional approach to perfor-

mance measurement relative to Southwest Airlines, and its correspond- ingly low levels of relational coordination.

Accountability versus Learning

As we have just seen, American and Southwest are on opposite ends of the spectrum with respect to performance measurement. At American, the purpose of performance measurement is accountability, often with a punitive twist. At Southwest, the purpose of performance measurement is to learn and improve over time. How did these airlines develop such different approaches?

American Airlines was designed at its inception to operate in a tradi- tional military fashion, with managers given responsibility for perfor- mance, and frontline workers responsible only for following commands. When Robert Crandall became CEO in the 1980s, he tried to make American Airlines less bureaucratic by increasing the accountability of employees at every level of the company. Using new systems for measur- ing and attributing performance outcomes, Crandall fostered a culture of accountability in which managers at every level would be held strictly accountable for their performance. Along with increased managerial accountability, he introduced employee participation throughout the

Avoid F inger Point ing—Measure Per formance Broadly 149

Chap 11 11/20/02 3:03 PM Page 149

C op

yr ig

ht ©

2 00

0. M

cG ra

w -H

ill T

ra de

. A ll

rig ht

s re

se rv

ed .

organization to push power and accountability down to frontline em- ployees. According to a station manager:

Under Crandall, we got the idea that everybody was accountable for results—not just top management. He started doing very tough reviews of budgets, for example, and increased the flow of company information to frontline workers by putting video machines in the ground workers’ ready room and starting a company paper.

People took on the idea of accountability. A lot of communication started to take place. . . . Crandall made these changes in response to deregulation, with an eye toward the competition.

Another station manager had a less positive perspective on this approach to accountability:

When Crandall came in the early 1980s, accountability was so new for us that it had a dramatic impact even though it was based on functional goals. Then people figured out how to game it, and headquarters kept tightening the screws.

In her view, the new measures were not conducive to cross-functional coordination. Accountability was perceived to be punitive, which tended to make people focus on pleasing their superiors, and fear making mis- takes or giving power to subordinates. Also, accountability at each level was pinned to individuals or functions rather than to the larger process, making people tend to look out for themselves and avoid recrimination, rather than focusing on their shared goals—on-time departures and sat- isfied customers. According to a Boston ramp manager:

The hardest part about the restructuring has been changing the way I man- age. I have grown a lot as a manager over this past year, but it has not been easy. . . . It is scary to delegate, especially at American, where there is a very strong company culture toward accountability. This is fine, but the penalties that go along with that accountability make people afraid to take risks— afraid to let go of their control.

To push decision making and responsibility down, you have to be willing to let others make mistakes and learn from them. [But for this] you need the full support of your boss.

150 T E N S O U T H W E S T P R A C T I C E S

Chap 11 11/20/02 3:03 PM Page 150

C op

yr ig

ht ©

2 00

0. M

cG ra

w -H

ill T

ra de

. A ll

rig ht

s re

se rv

ed .

This support from one’s boss did not seem to be present. Managers felt unable to play a supportive role for their direct reports because they were being judged critically by the managers above them, and so on. As part of the focus on accountability at American, station managers were given annual performance standards, called Minimum Acceptable Per- formance Standards (MAPS), for on-time departures, baggage handling, and customer complaints. A station manager complained:

I am harassed on a daily basis. Headquarters has a performance analysis department that is looking at my MAPS every day, analyzing the station’s performance.

A headquarters manager acknowledged the concern: “Failure to meet MAPS is perceived to result in punitive action.”

Several managers interviewed at American Airlines commented that there was a split between the stations and headquarters, and that the information flow between them was based largely on the numbers. According to a manager of human resources:

In this company, accountability is statistical. Managers are not judged on how well they delegate. They are judged by their results. The station man- ager is judged on the numbers and not on how he got them. He could have used a club for all it matters to his rewards.

An employee relations manager concurred:

All that matters is the numbers; how you achieve them is secondary. This is part of the culture of fear.

The reaction to the split between field and headquarters on the part of some station managers was bitterness. “Better communication clearly matters at the station level, but it doesn’t make a bit of difference what they do at headquarters,” according to a station manager. Managers per- ceived that they were judged strictly on the numbers. Information in the form of numbers went from stations to headquarters for evaluation, but there was little discussion—and little learning. At the same time, man- agers transmitted to frontline workers the pressures they perceived from headquarters, and at times used methods to achieve goals that created

Avoid F inger Point ing—Measure Per formance Broadly 151

Chap 11 11/20/02 3:03 PM Page 151

C op

yr ig

ht ©

2 00

0. M

cG ra

w -H

ill T

ra de

. A ll

rig ht

s re

se rv

ed .

resentment. Employees were well aware of their managers’ performance evaluation system, and how it affected them. A customer service agent explained, with bitterness:

Here you only care about delays. Otherwise the little report card won’t look good that week. The ultimate goal is not the customers, it’s the report card.

A Different Approach to Performance Measurement at Southwest

Like American, Southwest had adopted a formal hierarchical organiza- tional structure at its inception. But given the airline’s strategic focus on quick turnarounds, coordination of the flight departure process by front- line employees was emphasized from the start. Jim Wimberly, executive vice president of operations, explained:

It’s not, folks in Engineering do this, folks in Marketing do that. We recog- nized early on that it would only work if we worked as teams.

Southwest’s approach to accountability can be seen not only at the front line, with the team delay, but also in the relationship between field and headquarters. In stark contrast to what was seen at other airlines, particularly American, there is a two-way flow of information between field and headquarters at Southwest and appears to be a great deal of learning. When station managers at Southwest were asked how their own performance was assessed, they were quite vague about it. “I don’t know,” was one typical response, given with a laugh. “I’ll hear about it if I’m not doing a good job. I get free rein if I do OK.” “It is watched, but there is no fear factor,” said the Chicago station manager. “Everybody here is a self-motivator.” According to the Phoenix station manager:

I know what the relationship [between headquarters and the station] is usu- ally like because I worked at Eastern for 20 years. It’s usually an entrenched bureaucracy between the station manager and headquarters. It’s nothing like that here.

His assistant station manager explained:

152 T E N S O U T H W E S T P R A C T I C E S

Chap 11 11/20/02 3:03 PM Page 152

C op

yr ig

ht ©

2 00

0. M

cG ra

w -H

ill T

ra de

. A ll

rig ht

s re

se rv

ed .

Each station is like an entrepreneur. We do what we think is right, and talk directly to our executive vice presidents and Herb. They are just a phone call away. If they question something we did today, they will call tomorrow.

In contrast to the oft-heard claim that American’s system of perfor- mance measurement did not take into account how managers achieved their results, managers at Southwest claimed that they were judged by how they achieved their results. According to the Chicago station manager:

My director will look at overtime costs and on-time performance. But the biggest thing he’ll look at is morale. He believes that morale affects the other things. If you are not treating people right, the other things will get you. They’ll get you back in other ways.

At American, as we saw, the relationship was more hands-off, based on a much different flow of information. There was a perception by managers that they were judged strictly on the numbers. “Headquarters doesn’t care how you get the numbers, just that you get them,” was repeated time and again. Information in the form of numbers went from stations to headquarters for evaluation, but there was little discussion and little learning.

Summing Up

This chapter shows that cross-functional performance measurement improves coordination through its positive impact on relationships of shared goals, shared knowledge, and mutual respect, resulting in better performance. However, this approach to performance measurement flies in the face of classical organizational theory. Traditional management the- orists including Max Weber and Chester Barnard believed that functional accountability was the most effective way to achieve control.4 If you could clearly delineate someone’s realm of responsibility, they believed, then you could clearly measure how they are doing, and reward or punish them for those results. Functional accountability was also a means to avoid overload. Herbert Simon argued that human beings have a limited scope of atten- tion within which they can act in a rational way.5 Simon recommended functional accountability as a way to focus employee attention on a limited

Avoid F inger Point ing—Measure Per formance Broadly 153

Chap 11 11/20/02 3:03 PM Page 153

C op

yr ig

ht ©

2 00

0. M

cG ra

w -H

ill T

ra de

. A ll

rig ht

s re

se rv

ed .

set of responsibilities. He recognized the risks of functional accountability, particularly that employees would tend to focus on functional goals at the expense of organizational goals and might therefore fail to cooperate across functional boundaries. Still, he believed, the gains achieved from functional accountability outweighed these risks.

As we have just seen, however, trying to achieve control through functional accountability can seriously undermine information sharing and learning. Preoccupation with functional accountability leads to blaming, which in turn causes information to be distorted or to go underground.6 Because organizations need to use mistakes as a basis for learning, they should not rely on functional performance evaluation. Traditional measurement systems are flawed because they orient em- ployee attention toward functional rather than cross-functional out- comes and because they provide inadequate information for learning.7

To orient employees toward cross-functional outcomes and to provide more useful feedback about what to do, cross-functional performance measures should be used to supplement traditional functional measure- ment systems. These insights are supported by much of the recent orga- nizational literature.8 What we learned in this chapter, in addition, is that traditional performance measurement systems undermine performance because they weaken relationships of shared goals, shared knowledge, and mutual respect among those whose cooperation is critical for achieving the organization’s performance objectives.

154 T E N S O U T H W E S T P R A C T I C E S

Chap 11 11/20/02 3:03 PM Page 154

C op

yr ig

ht ©

2 00

0. M

cG ra

w -H

ill T

ra de

. A ll

rig ht

s re

se rv

ed .