Table of Contents
Go Back
Copyright 2001, by The Regents of the University of California.
Reprinted from the California Management Review, Vol. 43, No. 4.
by permission of The Regents.
Tradeoffs in Responses
to Work Pressure in
the Service Industry
Rogelio Oliva
ncreasing class size throughout the public education system and teachers’
burnout; rising airline load factors and airport utilization rates reaching
historical highs; long waiting times for emergency care and shorter hospital
stays after surgery—these are all symptoms of systematic imbalances in
demand and supply in service-providing organizations. The service sector repre-
sents over 70 percent of the U.S. economy and provides almost three out of
every four jobs in America. Studies consistently report that the main determi-
nant of customer satisfaction, even when purchasing sophisticated products, is
the level of service obtained from the supplier. Furthermore, managers of service
firms are also service consumers, and they surely appreciate the frustration of
poor or unfair treatment in a service transaction. It seems paradoxical that ser-
vices are recognized as critically important to our economy and yet episodes of
poor service proliferate. The complaint of poor services has regularly been
picked up by the popular press over the last twenty years. Even during a boom-
ing economy, we still do not seem to get services right: the American Customer
Satisfaction Index for services fell in 1999 to 69.4, down five percentage points
from its 1994 value.”
While imbalances in supply and demand eventually translate into con-
sequences for customers (long waiting times or reduced attention from service
personnel) and for the profitability of service enterprises (excess capacity or
reduced revenues due to unsatisfied customers), they first manifest as work
pressure for the servers. Service personnel perceive work pressure as the differ-
ence between the amount of work that can feasibly be done and the amount of
Support for this research has been provided by the Inventing the O rganizations of the 21st Century
Initiative at the MIT Sloan School of Management and the Division of Research at the Harvard Business
School.
26 CALIFORNIA MANAGEMENT REVIEW VOL.43,NO.4 SUMMER 2001
Tradeoffs in Responses to W ork Pressure in the Service Industry
work that needs to be performed. Under work pressure, service personnel strug-
gle to keep a balance between the flows of incoming and outgoing orders while
maintaining reasonable working hours and sustaining service quality. How a
service organization responds to work pressure is a critical determinant of ser-
vice quality, employee satisfaction, and the overall profitability of the service
firm. Consider the following excerpts from interviews with nurses of a presti-
gious teaching hospital in the midst of consolidating functions after a recent
merger:
“It’s not the desirable place it used to be to work. . . . You work very, very hard
while you’re here. . . . Nothing, absolutely nothing, laid back about this job. It’s
very high pressure, very high stress. . . . This is not going to be a place that I’m
going to stay long term.”
“The nurses are feeling more rushed in getting patients through here. . . . | think
patients are unhappy, and the nurses are unhappy because they're not able to
give time and care to a patient that they wanted to. And | think probably things
are being missed. | think patient care has suffered.”>
Since services are produced and consumed instantaneously, service orga-
nizations are particularly vulnerable to these imbalances in supply and demand.
That is, server and consumer have to be available at the same time for the ser-
vice transaction to take place. This simultaneity leaves no room for a “finished
goods inventory” to buffer the service delivery system from variations in
demand. The problem of balancing supply and demand in services, however,
is not simply a matter of absorbing short-term variations in customer orders; it
persists for two reasons. First, growth results in an imbalance as investments to
increase capacity struggle to keep pace with increasing demand. Over the last
fifty years, the service sector has consistently been the fastest growing sector in
the economy, and the situation is becoming critical as the U.S. economy is expe-
riencing extremely tight labor markets. Nationwide unemployment stands at
about 3.9 percent, and in some urban areas the figure is even lower. The U.S.
Bureau of Labor Statistics estimates that, if current trends continue, by 2005 the
total labor force will only be 2 percent higher than the total number of jobs in
the U.S. economy.*
A second force sustaining systematic imbalances in supply and demand is
the fact that service-sector productivity is improving more slowly than manufac-
turing productivity, resulting in increasing costs for the service sector.> Increasing
operating costs translate into financial pressures that drive service organizations
to process improvement and cost containment initiatives to seek efficiency gains.
This continuous search for productivity gains might be the only way for service
organizations to remain viable in a highly competitive environment. These ini-
tiatives, however, also ensure that the service organizations operate close to the
balance point between supply and demand, thus increasing the risk of tempo-
rary imbalances or even systematic underinvestment in service capacity.
CALIFORNIA MANAGEMENT REVIEW VOL.43,NO.4 SUMMER 2001 ve
Tradeoffs in Responses to W ork Pressure in the Service Industry
Since work pressure has a direct impact on service quality, employee sat-
isfaction, and overall profitability, it is important to understand how a service
organization responds to changes in work pressure and why it responds the way
it does. More importantly, we must comprehend the consequences from each of
these responses and identify what management can do about it.
Services Require an Integrative Approach
Services delivery processes differ from product manufacturing as the
servers (employees) and the elements being processed (customers) are
humans—with psychological attributes, perceptions, and expectations. Further-
more, services are produced in front of customers and often with direct collabo-
ration from them, thus bringing employees and customers physically and
psychologically close, Customers’ perception of the service experience is not
only affected by the conditions under which the service is delivered, but also by
the employees’ attitudes towards the customer. Similarly, employees’ attitudes
towards and perceptions of their job are influenced by customers’ attitudes
towards the service. This co-evolution of perceptions and expectations is further
confounded by the fact that services are intangible, thus making it difficult to
assess customer requirements and to fix an objective service standard. Clearly,
the study of services requires an interdisciplinary approach; an integrated under-
standing of the organizational and behavioral components of the social systems
that produce and consume the service, as well as the physical and technological
characteristics of the service delivery system.
This research reported here uses the system dynamics method to explore
different responses to work pressure. | created a formal simulation mode! that
captures the structural characteristics of the service delivery process, manage-
ment’s and employees’ decision-making processes, and the formation of expecta-
tions for customers and employees.® | found that the major recurring problems
observed in service industry—erosion of service quality, high turnover, and low
profitability’7—can be explained by the organization's response to changes in
work pressure. That is, the manner in which a service firm responds to work
pressure determines whether the system will disappoint customers, employees,
or shareholders. Furthermore, this research suggests ways to identify the struc-
tural characteristics that determine the preferred response in a given circum-
stance, thus making it possible to design high-leverage interventions and
policies. Although quality, costs, and employee satisfaction are normally per-
ceived as tradeoffs, | found that successful policies manage simultaneously to
delight customers, employees, and shareholders.
Responses to Work Pressure
Work pressure in a service setting typically manifests itself through a
backlog of customers waiting to be processed. The service backlog is the accumu-
lated difference of the incoming customer orders and the order fulfillment rates.
28 CALIFORNIA MANAGEMENT REVIEW VOL.43,NO.4 SUMMER 2001
Tradeoffs in Responses to W ork Pressure in the Service Industry
FIGURE 1. Work Pressure: Imbalance in Supply and Demand
Customer Order
Orders =a Fulfillment
Bs vice a
& * Backlog
Desired
Delivery
Deley Service
\. ra Standard
- ot
+
Desired Arrows indicate the direction of
Service C apacity causalty. Signs (*+" and *-") indicate
+ Time Per the polarity of relationships.A “+"
Order ‘means that, all else equal, if the cause
+ increases (decreases), the effect
increases above (decreases below)
Service Work what it would otherwise have been.
Capacity ye Pressure Similarly, a” means that, all else
7 equal, if the cause increases
(decreases), the effect decreases
below (increases above) what it
would otherwise have been.
Examples of customer backlog include the line in front of a cashier, callers on-
hold in a calling center, or service requests waiting to be processed. From the
service backlog, it is possible to determine the capacity required to ensure that
service is provided within a certain time (delivery delay) and with a certain level
of quality (service standard). Work pressure is the difference between the
required capacity and the service capacity available (see Figure 1). If a target
delivery delay and a service standard must be maintained, then the larger the
service backlog, the higher the work pressure that the service providers feel.
The possible responses to work pressure in a service environment are
limited. An option is to limit the flow of incoming orders. This can be done in
different ways. The first possibility is to limit the holding capacity of the backlog.
For example, if the bar in a restaurant has no more places available, then poten-
tial customers can infer that the waiting time for the service (dinner) is so long
that it is not worth the wait. Similarly, some call centers report a busy signal for
the caller once the number of calls waiting to be processed exceeds certain limit.
A second alternative is to limit demand through price, or to use price incentives.
to balance the variations in customer demand. Once customers have entered the
service setting, however, short of customers deciding to opt out of the waiting
line (hanging up the phone or canceling their order), there are only three possi-
ble actions to reduce the backlog: Employees can work harder, employees can
reduce the time allocated to each customer, or management can decide to
increase service capacity.®
CALIFORNIA MANAGEMENT REVIEW VOL.43,NO.4 SUMMER 2001 29
Tradeoffs in Responses to W ork Pressure in the Service Industry
FIGURE 2. Responses to W ork Pressure
Customer Order
Orders = Fulfillment
2 vice ?
S rs P) Backlog
Delvery Comer Cutting
elay Service
Standard
» * KO
+
Desired
Service C apacity
+ Time Per
Order
* - Overtime
work + Work
Service Pressure
Capacity” 53 ——— Intensity
Capacity
Adjustment
Delays separating the cause and the
effect are indicated in the diagrams by
delay boxes. Balancing loop polarity
(denoted by B in the loop identifier)
indicates a regulating (negative)
feedback loop.
The first option for service providers experiencing work pressure is to
increase their work intensity (WI), i.e., to work harder. This normally translates
into providers’ reducing the number and length of breaks that they take during
the day. If this is not enough, service providers will eventually incur overtime.
Asall the other responses to work pressure, increasing the WI creates a balanc-
ing mechanism that attempts to maintain work pressure within certain limits.
A higher work intensity increases the order fulfillment rate, thus reducing the
service backlog and eliminating the work pressure (see the Overtime Loop in
Figure 2). Under extreme work pressure, management might decide to reduce
all training, planning, and improvement efforts and encourage service personnel
to allocate all time available to order processing. This is typical behavior for orga-
nizations facing major deadlines.
A second response to changes in work pressure is for service providers to
adjust the time per order (TPO), i.e., the time allocated to each customer order.
Although the TPO is normally determined by the firm’s service standard, in
times of high work pressure service personnel can attempt to process each cus-
tomer faster. Speeding up service transactions might be as simple as reducing
the time spent in pleasantries with the customer, but it might also extend to
eliminating post-service documentation or some core aspects of the service
delivery process, i.e., “cutting corners” from the full-service transaction. Given
a constant service capacity, reducing the time per order effectively increases the
order fulfillment rate, thus reducing the existing backlog and eventually elimi-
nating the work pressure (see the Corner-Cutting Loop in Figure 2).
30 CALIFORNIA MANAGEMENT REVIEW VOL.43,NO.4 SUMMER 2001
Tradeoffs in Responses to W ork Pressure in the Service Industry
FIGURE 3. Consequences of Employees’ Responses to W ork Pressure
Customer Order
Orders Fulfillment
5 Service
= > Backlog
Desired 62
Delivery
Corner Cutting
Dek ,
a Service -
Standard
\ Wee t Fatigue
Service Capacity Goal Osi ne Per (oy)
Order /Burnoy
A (1
+ 0 Reinforcing loop polarity (denoted
Work polly by R in the loop identifier)
Pressure + Work
indicates a setf-einforcing
Intensity (positive) feedback process.
The third response to changes in work pressure is to invest in service
capacity (SC). Higher service capacity increases the order fulfillment rate, which
in turn reduces the service backlog, the required service capacity, and work pres-
sure (see the Capacity Adjustment Loop in Figure 2). Increasing service capacity
can be accomplished by increasing the number of employees, investing in equip-
ment and technology, or redesigning processes. Regardless of the selected alter-
native, adjustments of service capacity are not instantaneous and, in some cases,
the introduction of new resources (computers or employees without experience)
or the adoption of new technologies is disruptive in the short term.°
Performance Traps
Since increasing service capacity usually involves higher expenditures
and longer delays than the other two alternatives, companies usually opt to
allow their service personnel to absorb small variations in work pressure. Unfor-
tunately, using employees’ responses to work pressure (WI and TPO) over an
extended period has some undesirable consequences. First, sustained periods
of high WI cause fatigue that eventually reduces the productivity of the service
providers. Reduced productivity of service providers leads to a lower order ful-
fillment rate that, everything else being equal, translates into a higher service
backlog, further increasing work pressure and forcing service providers to work
even harder (see Burnout Loop in Figure 3).?° The second unintended conse-
quence of using employees’ responses to work pressure is subtler. Because of a
lack of an objective service standard, extended periods of reduced TPO tend to
modify the service standard as employees, managers, and customers adjust
their service expectations to past performance." With lower service standards,
the time allocated per order is reduced, thus reducing the service standard
even further (see Goal Erosion Loop in Figure 3). Note that these unintended
CALIFORNIA MANAGEMENT REVIEW VOL.43,NO.4 SUMMER 2001 31
Tradeoffs in Responses to W ork Pressure in the Service Industry
consequences are reinforcing traps (bad performance yields even worse results)
that erode the work quality for employees or the overall service quality of the
firm.
How should organization respond to changes in work pressure? The
choice is not obvious. Only increasing service capacity avoids the performance
traps of employee burnout or erosion of service standards. Adjustments of
service capacity, however, are not instantaneous and directly affect the cost
structure of the service firm. To a certain extent, it seems desirable to use the
flexibility and immediacy of the employees’ responses to deal with short vari-
ations in work pressure. While the individual effects of these responses to work
pressure are simple to understand, it is more difficult to comprehend the trade-
offs presented by the interaction of these responses and to assess their long-term
consequences. A well-calibrated and tested simulation model can help in under-
standing these dynamic tradeoffs.
Responses of a Lending Center
In 1990, a major retail bank in the UK sought to cut costs by moving
back-office operations from branches to centralized processing centers in more
affordable locations. Under this scheme, a Lending Center serves as a back-office
for managing personal loans, credit cards, and small business accounts. Work
arrives to the Lending Center by telephone (customer inquiries), mail (customer
requests and communications with branches), and daily computer-generated
reports identifying problematic accounts that require immediate action (such
as overdrafts and missing payments). M ost requests produce either a letter or a
telephone conversation with the customer. Tasks are monitored against standard
processing times, and there is a clear expectation that all tasks will be processed
within 24 hours of their arrival.
While the strategy to centralize and standardize the back-office operations
had improved the productivity of the lending officers and the quality of the
lending book,"? employees with direct customer contact had some concerns
about the level of service provided. When asked how they felt the Lending Cen-
ters were working for the bank, here is what two lending officers had to say:
“The feedback you get back [from the customer] is ‘I’m dehumanized, | just
became a number. | can no longer talk to you as a person, you just treat me as
anumber.’. . . we have lost the customers along the way.”
“| think it has been very effective in actually improving the lending book because
of the monitoring system. Now that we've seen [the reduction of the risk index]
through, because of the various pressures on us, we are going to be asked to be
more proactive in selling. . .. We just don’t have the relationship basis to sell
effectively. The customers have said that they become a number; and in a way
they have. .. . It is difficult to sell that way.”
32 CALIFORNIA MANAGEMENT REVIEW VOL.43,NO.4 SUMMER 2001
Tradeoffs in Responses to W ork Pressure in the Service Industry
FIGURE 4. Lending 0 fficers’ Response to W ork Pressure
11
Effect of W P on
Work Intensity
105 4 aN
Effect of WP on
Time Per Order —™”
ee Pee eee eet
os 4 i
09 T T t T T T
-0.15 -01 -0.05 0 0.05 01 0.15
Work Pressure
Fieldwork was done at a Lending Center supporting 20 branches in the
West End region of London. Data collected for the study included: time series
for the operational metrics of the service center; interviews with employees,
their managers, and staff inside and outside the Lending Center; direct observa-
tion; and archival data such as procedure manuals and training materials. | used
these data to specify the decision rules of employees and managers in the model,
as well as the physical constraints of the service delivery process. The adequacy
of the mode! was tested by evaluating the rationality of the policies estimated for
decision makers and the model's ability to replicate historical behavior for time
per order, work intensity, order fulfillment, and multiple metrics of service
capacity.
Analyzing the operational performance, | found that, under pressure to
increase output, lending officers are much more willing to cut corners (reduce
TPO) and only reluctantly work longer hours (increase WI). Figure 4 shows the
lending officers’ estimated response to work pressure. In the absence of work
pressure (WP=0), lending officers work their regular hours and allocate to each
customer order the effort dictated by the service standard—effect of work pres-
sure on WI and TPO is equal to one (neutral). If work pressure rises above its
neutral level (0), employees simultaneously increase their work intensity and
reduce the time per order. Their response on cutting corners, however, is almost
twice as aggressive as their increase of work intensity. Although in interviews
and surveys employees claimed a deep concern for the “standard of customer
service” of the 15 loan officers interviewed, all but one admitted to reducing
their effort to document transactions and sell additional products in times of
high work pressure. Note that consistent with the emphasis the monitoring
CALIFORNIA MANAGEMENT REVIEW VOL.43,NO.4 SUMMER 2001 33
Tradeoffs in Responses to W ork Pressure in the Service Industry
FIGURE 5. Stack Responses to a 10% Increase in Demand
FA DP DPD PO
Throughput
110 4 Le
8 Permanent Erosion of Service Standard
i
° 105-7 bes:
ey
Fs (Sc) Capacity Expansion “ /
1.00 bs
T T T E T \ T T T
system places on processing customer orders the same day they arrive, employ-
ees’ response to work pressure is at all times enough to compensate for the
change in work pressure. For example, a5 percent increase in work pressure
is handled through a 1.8 percent increase in work intensity and a 3.2 percent
reduction of time per order.
Consistent with the interviews with loan officers, simulations showed a 4
percent reduction in the service standard—measured in time per order—for the
period when data were available. The erosion persisted even when the model
was initiated in equilibrium, ruling out transient effects from the opening of the
Lending Center. Furthermore, extended models simulations showed that the
erosion of service continued well beyond the point where service capacity and
service demand had reached equilibrium (no work pressure).
To isolate how the interactions among the three responses to work pres-
sure—increasing WI, reducing TPO, and increasing SC—generate the erosion of
the service standard, | tested the model without the random variations of cus-
tomer orders. Initially, the service center is balanced: employees are not working
overtime and customers receive the expected level of service. At week 10, the
simulation introduced and sustained, a 10 percent increase in customer orders.
Since the Lending Center has a commitment to process all orders within 24
hours (i.e., constant delivery delay), the excess orders require the Lending Cen-
ter to increase its throughput by 10 percent. Figure 5 shows the contribution to
throughput achieved by each of the responses to work pressure, as well as the
erosion of the service standard. For example, five weeks after the introduction of
the incremental demand (week 15), service capacity has not yet been expanded,
and the service standard has already eroded 2 percent because of aggressive
corner cutting during the initial weeks of work pressure. The remainder of the
34 CALIFORNIA MANAGEMENT REVIEW VOL.43,NO.4 SUMMER 2001
Tradeoffs in Responses to W ork Pressure in the Service Industry
required increase in throughput is achieved through a 3 percent increase in
overtime, and a5 percent reduction of time per order beyond the current
standard.
The combination of responses is effective in immediately increasing
throughput—at all times the combination of responses adds up to the required
10 percent increase. However, the timing and strength of these responses differ
substantially. As explained above, the initial reduction of TPO is almost twice as
aggressive as the increase in WI. Furthermore, whereas employees’ responses to
work pressure (TPO and WI) are essentially instantaneous, the adjustment of
service capacity is slow, peaking after 25 weeks. There are several reasons for
the lag in the response to adjust SC. First, although performance metrics are
available on a weekly basis, these metrics are summarized and analyzed only at
the end of the month. Next, to smooth out the high frequency variations in cus-
tomer orders, managers adjust their estimate of required service capacity with
an average lag of 4 months. The delay in adjusting authorized labor achieves
its purpose of filtering out variations in customer orders and is consistent with
management's imperative to control costs. Once labor has been authorized, it
takes on average seven months for the hiring process to bring a new employee
into the Lending Center. Finally, rookies are on average 35 percent as productive
as experienced personnel, with an average delay of 12 months before becoming
fully productive. The combination of cautious hiring policies, hiring delays, and
long training requirements cause service capacity to react slowly to changes in
demand. The consequence of the slow adjustment of SC is that temporary varia-
tionsin work pressure must be accommodated through either cutting corners or
working overtime.
Model simulations with historical variation patterns showed that, as
expected, employees absorb small increases in work pressure arising from varia-
tionsin demand and absenteeism by reducing TPO and increasing WI. Although
reducing TPO enabled an immediate increase in throughput, it also resulted in
the erosion of the internal service standard (the Goal Erosion Loop in Figure 6).
Furthermore, in the absence of direct, reliable, and trusted measurements of
customer satisfaction, management interprets the reduction in TPO and service
standard as productivity gains, and adjusts the required service capacity accord-
ingly. The adjustment of desired service capacity through the erosion of the ser-
vice standard should eventually eliminate work pressure and bring the system
to equilibrium (the False Learning Loop in Figure 6).1> However, the reduction
of desired service capacity also translates, with a delay, into a reduction of the
actual service capacity. Lower service capacity further increases work pressure
on the service delivery personnel, who in turn reduce TPO, thus locking the
system into a vicious cycle of eroding standards and diminishing service capacity
(the Death Spiral Loop in Figure 6).
The relative strength and timing of the responses (TPO>WI>SC) accounts
for the observed erosion of service standards. By the time hiring reacts to the
changes in customer orders and new employees are trained, the required service
CALIFORNIA MANAGEMENT REVIEW VOL.43,NO.4 SUMMER 2001 35
Tradeoffs in Responses to W ork Pressure in the Service Industry
FIGURE 6. Erosion of Service Standard
Customer Order
Orders Te Fulfillment
oF & Backlog
Desired @}
pehey Corner Cutting
id Service
\ + am Standard “f -
Desired @} oe
Service Capacity Goal Erosion
Gah Time Per
Order
False Learning +
: 7
Work
Spiral Pressure
—*
Capacity = @s
Capacity
Adjustment
capacity has eroded with the service standard. Regressions showed that the
reduction of the service standard experienced during the period for which data
were available translated into a 50 percent reduction in expected sales. As large
as lost sales can be, they still underestimate the hidden costs of a low service
standard, since high work pressure also translates into errors in documentation
and higher rework rates.
Policy Analysis
Once the reasons for quality erosion were understood, | used the model
to develop policy recommendations. Specifically, we were interested in develop-
ing policies to maintain service quality without compromising the organization's
ability to respond to demand fluctuations and without carrying excess capacity.
There are only three ways in which the Lending Center's response profile for
changes in work pressure (Figure 5) can be modified while still maintaining
full responsiveness: expediting the adjustment of service capacity, reducing the
effect of work pressure on time per order, or reducing the rate at which the ser-
vice standard erodes. These three changes correspond to moving “up” each of
the lines separating the responses in Figure 5. Following is a list of possible
implementations of these strategies.
= Expediting the Adjustment of Capacity— Since the erosion of the internal ser-
vice standard occurs when work pressure is high, one obvious policy is
to ensure that service capacity is acquired before the standard can erode.
Capacity expansion can be expedited by reducing the delays governing
36 CALIFORNIA MANAGEMENT REVIEW VOL.43,NO.4 SUMMER 2001
Tradeoffs in Responses to W ork Pressure in the Service Industry
the capacity adjustment process and creating a more responsive hiring
process. A second option is to reduce the time it takes rookies to become
fully productive. Accelerated learning rates can be achieved either by
hiring employees with greater initial effectiveness or by increasing the
intensity of their training programs. Unfortunately, the option to acceler-
ate learning curves is rarely available in high-contact services that require
job-specific knowledge.
= Reducing the Effect of Work Pressure on Time Per Order— The positive feedback
driving the erosion of the service standard is triggered by cuts in time per
order caused by high work pressure. Reducing employees’ willingness to
cut corners should weaken the Goal Erosion Loop and slow the decline
of the service standard. This normally requires management to become
aware of the implications of poor service—lost sales, rework, and cus-
tomer defections—and to persuade employees to avoiding them; even
at the cost of slower transactions. Of course, if the time spent with cus-
tomers were completely unaffected by work pressure, there could be no
quality erosion. Such a rigid policy, however, is unfeasible in high-contact
services where employees have considerable autonomy in deciding how
they respond to each customer. A more realistic policy should aim to dis-
tribute employee responses to work pressure evenly between corner cut-
ting and overtime. This can be done by reducing the flexibility of the
service encounter (through process standardization and documentation)
or by increasing the relative attractiveness of overtime (by creating high
empathy with customers or increasing overtime compensation).
= Redudng the Erosion Rate for the Service Standard— The intent of the previous
two policies is to reduce the usage of TPO as a coping strategy for changes
in work pressure. Under unavoidable variations of work pressure, how-
ever, maintaining service quality requires explicit pressure to prevent the
erosion of the standard. An external norm for service quality may provide
sufficient counterpressure to halt standard erosion. In some industries,
such external norms are developed as part of the professional training
of service providers. More often, management must take an active role
in forming the service standard by articulating clear and consistent expec-
tations for service quality and monitoring performance against them.
How much should each of these levers be pushed in order to maintain
service quality and full responsiveness to variations in demand? Two separate
analyses are required to answer this question. First, it is necessary to determine
how successful a policy isin achieving its stated goal. Because of non-linear rela-
tionships and multiple feedback mechanisms interacting in a service setting, it is
difficult to predict the effectiveness of a proposed policy and the unintended
consequences that it might generate. Experimentation with a computer simula-
tion model is an effective way to assess the impact of each proposed policy and
its interactions with other elements of the system. For example, despite the
apparent benefits of reducing the time for employees to become fully trained,
CALIFORNIA MANAGEMENT REVIEW VOL.43,NO.4 SUMMER 2001 37
Tradeoffs in Responses to W ork Pressure in the Service Industry
FIGURE 7. Stack Responses to a 10% Increase in Demand after Implementation
of Policy Recommendations
Fa ee ee A A SY
Throughput
110 4 =
Permanent Erosion of Service Standard
da
&. aise (WI) Overtime (TPO) lL
Temporary
4 Corner Cutting
-
= (SC) Capacity Expansion
1.00 =
a a a |
| found that this policy had limited impact in the Lending Center since on aver-
age only 6 percent of its employees are inexperienced at any given time.
Once a policy is deemed effective, it is necessary to assess how easy it
would be to implement the policy in the real system. Policy changes require
structural modifications—i.e., changes in norms, processes, and decision making.
The feasibility of structural changes in a particular situation is a function of the
existing conditions and the cost to implement the changes. For example, in the
case of the Lending Center it was desirable to reduce the delays of the labor
authorization and hiring processes—a combined delay of 11 months. However,
the existing set of relationships between the Lending Center and the district
office of human resources made it feasible to think only of a 50 percent reduc-
tion in these delays.
In the case of the Lending Center, no one single policy was successful
in stopping the erosion of the quality standards and a combination of the
three policies described above was recommended. Figure 7 shows the modified
response profile for the Lending Center after the three policies were imple-
mented simultaneously. Notice that WI takes a much larger share of the required
throughput, as quality norms prevent the extended use of TPO. Faster capacity
acquisition and more emphasis on overtime did reduce the effect of work pres-
sure on time per order and the service standard. However, in order to raise the
standard after its initial erosion, it was necessary to introduce managerial pres-
sure that would drive employees to deliver at levels above their current internal
standard. In general, | found that policies that maintain a balance among the
different responses are more successful at maintaining throughput and avoiding
the performance traps. That is, if the excess work pressure is distributed among
38 CALIFORNIA MANAGEMENT REVIEW VOL.43,NO.4 SUMMER 2001
Tradeoffs in Responses to W ork Pressure in the Service Industry
multiple responses, the unintended consequences from all the responses are
minimized.
Structural Determinants of Response Flexibility
The relative intensity of the responses to work pressure in the Lending
Center are determined by the structural characteristics of its service delivery
system, specifically the need to customize service transactions and the delays in
hiring and developing employee skills. Customization inhibits standardization of
the service delivery process, thus allowing service employees to reduce service
scope in response to work pressure. A significant but slow learning curve
reduces the speed at which service capacity can be acquired. The specifics cer-
tainly vary from industry to industry—for example, service settings with high
professional standards will have stronger quality pressure and slower erosion of
standards. However, given the broad prevalence of on-the-job training, delays in
capacity acquisition, and the intangibility of services, the structure that can lead
to erosion of service quality is likely to be common throughout the service
sector.
Erosion of service quality, however, is not the only possible outcome from
work pressure. Other service settings have a different set of dominant structural
characteristics that affect the strength of the responses to work pressure. For
example, the use of the reduction of TPO as a way to deal with work pressure
is limited if the service delivery process has been standardized, or if professional
standards constrain the customer-facing personnel to provide a certain service
level. The WI response is limited by the working hours in the service setting and
the willingness and incentives that employees have for working overtime.
Finally, the responsiveness of changes in SC is limited by the amount of training
required, the speed at which additional service capacity can be acquired, and the
managerial policies in place. A detailed list of the main constraints on each of
the response mechanisms is presented in Table 1. Each of the limiting factors
relates to a structural characteristic of the service setting.
Once the structural characteristics of the service setting are mapped
into the basic responses to work pressure, it is possible to identify the relative
strength of the responses to work pressure that could be expected from those
structural components. For example, in a service with a standardized (rigid)
delivery process (low TPO flexibility) and relative short training requirements
(high SC flexibility)—e.g., a fast-food restaurant—the relative strength of the
responses to work pressure that could be expected is: first, increase of WI, then
increase SC, and, probably very weakly, a reduction in TPO. Table 2 presents
examples of service settings with stylized structural characteristics and their
expected response preference to changes in work pressure. It is worth noting
that services with different structural characteristics could show the same rank-
ing of responses. For example, a capital-intensive firm with fixed standards of
service—e.g., a utility—that is forced to use equipment beyond the natural
CALIFORNIA MANAGEMENT REVIEW VOL.43,NO.4 SUMMER 2001 39
Tradeoffs in Responses to W ork Pressure in the Service Industry
TABLE 1. Factors Limiting the Flexibility of Responses to W ork Pressure
Response Factors Limiting Flexibility of Response
Time Per Order Standardized Service Delivery Process (Low Customization)
Professional Q uality Standard (High Customization)
Quality Sensitive Customers
Good Information of Q uality Performance
Constraints on W orking Hours
High Customer Contact Time
Regulated W ork-Hours (Airline Pilots)
Employee's Lack of Empathy with Customers
Capital Intensity
LongTraining Requirements
Professional Certification
Long Hiring Delay
Financial Pressures
Work Intensity
Service Capacity
maintenance cycle to satisfy an increase of demand would have the same rank-
ing of responses as the fast-food restaurant (WI>SC>TPO).
Because the similar response rankings create similar dynamic behaviors,
regardless of the structural limitations causing the ranking, the relative strength
of the responses to work pressure can be used as a way to classify the dynamics
that characterize services. Such classification allows for a reduction of the
dimensional characteristics needed to differentiate service settings and creates
a direct linkage between the dominant structural characteristics of the service
setting, its behavior, and potential policy recommendations.
For instance, the high turnover observed in entry-level service jobs is not
surprising when exploring the structural characteristics of the settings where it is
normally encountered. J obs requiring few skills are characterized by standard-
ized service delivery process (low TPO flexibility), low face-to-face contact (high
WI flexibility), and very low margins that discourage the investment of service
capacity (low SC flexibility).* These structural characteristics suggest an over-
whelming preference of WI over the other potential responses to work pressure
(WI>>SC>TPO). By limiting the other two responses, through tight standardiza-
tion and austere capacity policies, management has—perhaps unintentionally—
selected to disappoint employees and use them as the escape valve for changes
in work pressure. It is, then, not surprising that employees would decide to leave
this type of job.
Similarly, the erosion of service quality and burnout reported by the
nurses in the teaching hospital (as described earlier) can be explained by the
dominant structural characteristics of the health care delivery system and the
responses that they generate. The attributes of health care practitioners, strict
professional quality standards and high empathy with customers, combined
with long training requirements and intensive capital investments to increase
40 CALIFORNIA MANAGEMENT REVIEW VOL.43,NO.4 SUMMER 2001
Tradeoffs in Responses to W ork Pressure in the Service Industry
TABLE 2. Response Flexibility: Examples, Structural Characteristics, and Expected
Response Preference
TPO Flexibil
High
sc
Flexibility
Low
Ticket Sales
* Standardized delivery Process
+ Limited W ork Hours
* Low Training Requirements
SC>TPO>WI
Fast-Food Restaurant
* Standardized Delivery Process
+ Low Face-to-Face Contact
* Low Training Requirements
WI>SC>TPO
Airline Pilots
+ Standardized Delivery Process
* Regulated W ork Hours
* LongTraining Requirements
SC>WI&TPO
Health Care
* Strict Professional Standards
* High Empathy with Customers
* LongTraining/C apital Intensive
WISTPO>SC
Low
High
WI Flexibility
HighTPO Flexibility
High
sc
Flexibility
Low
Claims Adjusting Process
* High Customization
* Limited W ork Hours
* LongTraining Requirements
TPO>WI>SC
Maintenance Crew
* High Customization
+ Low Face-to-Face Contact
+ LongTraining Requirements
TPO&WI>SC
Low
High
WI Flexibility
*The combination of high SC flexibility with high TPO flexibility is not feasible since high customization (TPO flexibility) implies long training
requirements (low SC flexibiity)
capacity, yield a response profile that would first tap into WI to compensate
work pressure. If high WI is sustained, health care practitioners will eventually
reduce the TPO and erode their service standard. This predicted response is con-
sistent with the employee burnout and erosion of the service standard reported
throughout the health care industry.
CALIFORNIA MANAGEMENT REVIEW VOL.43,NO.4 SUMMER 2001
41
Tradeoffs in Responses to W ork Pressure in the Service Industry
There are other reasons why a service firm might exhibit an undesired
pattern of behavior—for instance, high turnover might be driven by a tight labor
market. Although those environmental issues can be easily incorporated into the
analysis, the model focuses on endogenous explanations for the observed behav-
ior, i.e., most variables are internal to the service center and under managerial
control. The endogenous perspective, by making the tradeoffs among options
explicit, allows the model to be used to explore alternative intervention strate-
gies for improving performance. Finding an optimal mix of policies to achieve a
target quality level and responsiveness to demand still requires the specification
of a detailed cost function associated with each response and policy lever. How-
ever, by establishing the link between structural characteristics of the service
setting, flexibility of responses to work pressure, and undesired behaviors, the
framework presents a full description on how the feedback structure of a system
generates its behavior. Based on this framework, it is possible to design struc-
tural changes (high-leverage policies) to avoid the undesired consequences of
work pressure. Undesired behavior (high turnover, erosion of service quality,
low profitability) is, in most cases, a symptom of an unbalanced response to
changes in work pressure (work intensity, cutting corners, excess capacity).
Balancing the responses to work pressure should be a matter of decreasing the
flexibility of the abused response and increasing the flexibility of the untapped
responses. A direct understanding of the structural characteristics that determine
the flexibility of the each response makes the diagnosis and selection of leverage
points easier.
Notes
1. B. Tuchman, “The Decline of Quality,” N ew York Times Sunday Magazine, November
2, 1980, p. 38; J. Main, “Toward Service Without a Snarl,” Fortune, March 23,
1981, p. 58; M.R. Feinberg and A. Levenstein, “It’s Not My J ob, Man,” Wall Street
Journal, November 11, 1985; S. Koepp, “Why Is Service So Bad? Pul-eeze! Will
Somebody Help Me?” Time, February 2, 1987, p. 46; D. Brady, “Why Service
Stinks,” Business Week, October 23, 2000, pp. 118-128.
2. “American Customer Satisfaction Index,” American Society for Quality, 2000,
<http://acsi.asq.org/>.
3. D.B. Weinberg, “Why Are the Nurses Crying? Restructuring, Power, and Control
in an American Hospital,” unpublished Ph.D. dissertation, Harvard University,
Cambridge, MA, 2000.
4. D.S. Friedman, “Help Wanted,” The M Kinsey Quarterly, 1 (1998): 34-44.
5. W.Baumol, S.B. Blackman, and E. Wolf, Productivity and American Leadership (Cam-
bridge, MA: MIT Press, 1991).
6. For a full description and documentation of the model, see R. Oliva and J .D. Ster-
man, “Cutting Corners and Working Overtime: Quality Erosion in the Service
Industry,” Management Sdence, 47/7 (2001): 894-914.
7. For evidence on erosion of service quality, see D. Brady, “Why Service Stinks,”
Business Week, October 23, 2000, pp. 118-128; Oliva and Sterman, op cit. For ero-
sion of job scope and employee satisfaction, see L. Schlesinger and J. Heskett,
“Breaking the Cycle of Failure in Services,” Sloan Management Review, 32/3 (Spring
1991): 17-28. For low profitability in the service sector, see W.J. Baumol, S.A.B.
42 CALIFORNIA MANAGEMENT REVIEW VOL.43,NO.4 SUMMER 2001
Tradeoffs in Responses to W ork Pressure in the Service Industry
Blackman, and E.N. Wolff, “Unbalanced Growth Revisited: Asymptotic Stagnancy
and New Evidence,” American Economic Review, 75/4 (1985): 806-817; P.T. Harker,
“Introduction: Service-Sector Productivity—The M S/OR Challenge,” Interfaces,
25/3 (1995): 1-5.
8. Note that prioritizing customer orders—or subjecting them to a triage process—
does not help reduce work pressure. Through prioritization, it is possible to speed
up the response to an important customer or to a person requiring the service
more promptly, but the overall load over the service delivery system remains
unchanged.
9. For a discussion on the impact of improvement efforts in the order fulfillment
rate, see N.P. Repenning, L.J. Black, and P. Goncalves, “Why Good Process Some-
times Produce Bad Results,” published in this issue [California Management Review,
43/4 (Summer 2001)]; E. Keating, R. Oliva, N.P. Repenning, S. Rockart, and J .D.
Sterman, “Overcoming the Improvement Paradox,” European Management J ournal,
17/2 (1999): 120-134.
10. For a survey of evidence on the effects of overtime on labor productivity, see
H.R. Thomas, “Effects of Scheduled Overtime on Labor Productivity: A Literature
Review and Analysis,” PT19107, The Pennsylvania Transportation Institute, Penn-
sylvania State University, University Park, PA, 1990. Homer explores the burnout
mechanisms from a systems dynamics perspective. J.B. Homer, “Worker Burnout:
A Dynamic Model with Implications for Prevention and Control,” System Dynamics
Review, 1/1 (1985): 42-62.
11. SeeJ. March and H. Simon, Organizations (New York, NY: Wiley, 1958); R. Cyert
and J. March, A Behavioral Theory of the Firm (Englewood Cliffs, NJ : Prentice Hall,
1963). For empirical evidence, see T.K. Lant, “Aspiration Level Adaptation: An
Empirical Exploration,” Management Sdence, 38/5 (1992): 623-644.
12. Quality of the lending book is measured through a risk index calculated by
weighting the each account's risk grade by the loan amount. The lending centers
had yielded significant improvements of the book's risk index through centraliza-
tion and standardization of the lending criteria.
13. Notice in Figure 5 how the employee's responses disappear after the permanent
erosion of the service standard and service capacity reach an equilibrium and
work pressure returns to zero. The slight adjustment between capacity expansion
and the permanent erosion of service standard after week 48 is due to an over-
shoot of the hiring policy that results in negative work intensity. The system
slowly reaches equilibrium (through employee attrition) at the service level
attained when work pressure first reached equilibrium. In this particular test, the
simulated organization increased its throughput 10 percent by reducing the inter-
nal standard of customer service 5.4 percent and increasing service capacity 4.1
percent. Throughput (orders/week) is defined as service capacity (man-hours/
week), divided by time per order (man-hours/order). (1+0.041)/(1-0.054)=1.1.
14. For an account of an employee attempting to survive on low satisfaction, high
turnover services, see B. Ehrenreich, Nickel and Dimed: On (Not) Getting by in
America (New York, NY: Metropolitan Books, 2001). See also L. Schlesinger and
J. Heskett, “Breaking the Cycle of Failure in Services,” Sloan Management Review,
32/3 (Spring 1991): 17-28.
CALIFORNIA MANAGEMENT REVIEW VOL.43,NO.4 SUMMER 2001 43
Back to the Top
