Lean Service Machine

C:\DOCUME~1\Jonathan\LOCALS~1\Temp\ZEX0BCFA.htm 31/03/04

initial application so that he or she could physically assemble the policy and send it to

the adviser. It was possible that employees would spend extra time assembling

policies, leaving teams idle downstream and delaying the flow of application

processing. JPF split the receiving team in half, assigning some of the employees to

assemble the policies while the rest continued to receive applications. The change

required no additional space, equipment, or people, but it eliminated confusion on the

part of employees about what they should be doing when, and it substantially reduced

delays and waste.

Setting a Common Tempo. The team further smoothed out work flow by applying

the concept of “takt” time. Takt, the heartbeat of lean operations, is derived from the

German word for musical meter. It refers to pacing work according to customer

demand. The team knew that to satisfy demand, New Business needed to process ten

applications per hour; the takt time was therefore one application every six minutes. A

worker producing at a slower rate would leave the next person in line temporarily idle

and would ultimately prevent the group from meeting customer demand. The lean -

team members timed each work element of the model cell, such as retrieving an

application from an in box and performing the keystrokes required to get an

application into the system. They established a baseline time for each element by

determining how quickly an untrained person could do it, then challenged employees

to make improvements and create shorter baseline times. As workers found ways to

cut unnecessary tasks, the lean team determined the minimum number of employees

required for completion of all steps.

Balancing Loads. Lean production systems are designed to balance work evenly

among employees. Although workers appreciate the fairness of this practice, its

greater value is that it eliminates unnecessary delays. JPF had always allocated

incoming applications first by distribution channel and then alphabetically at each

stage. An application from a customer named Burns would be allocated to the A –C

team even if another team was idle. In the model cell, the alphabetical method was

replaced by sequential allocation, so that every team received the same number of

applications. This enabled work to flow smoothly from one fully utilized employee to

the next without unnecessary delays.

Segregating Complexity. Anyone who has stood in line at a bank while a single teller

assisted a customer with a lengthy transaction understands this principle. The key to

successfully segregating complexity is to cluster tasks of similar levels of difficulty into

separate groups with their own performance goals. Thus the model cell eventually

divided into two groups, one handling cases that did not require a physician statement

and the other handling those that did. Once the separation was made, the turnaround

time for cases not needing a doctor’s statement fell by more than 80%.

Posting Performance Results. Following one of lean manufacturing’s best practices,

JPF displayed the cell’s hourly productivity rates along with the company’s

expectations. The numbers were posted on large white boards so that all employees

could see when and where—and therefore why—performance was suffering. The team

also set aside an area by the boards so that employees could meet quickly to discuss

ways of solving performance problems that arose. Not surprisingly, the boards made a

few people uneasy. Some workers feared that the posted results would be used to

assign blame and punish low performers. But as employees grew accustomed to the

ubiquitous boards, the displays became rallying points for celebrating successes and

encouraging the team to set new performance records. Employees understood that

they would be evaluated on and rewarded for objective results they could track

themselves—rather than by their bosses’ subjective observations.

Setting Performance Goals

To implement lean production, a company typically must overhaul the way it measures

costs, speed, and quality. Indeed, managers often find that many of their company ’s

favorite metrics actually inhibit productivity. For example, if the performance of call -

Page 4 of 9

file://C:\DOCUME~1\Jonathan\LOCALS~1\Temp\ZEX0BCFA.htm 31/03/04

center reps is measured by the length of an average phone call, the center may get a

lot of repeat calls because customers’ concerns are not being resolved the first time.

To get around this kind of problem, practitioners of lean production follow an

important principle: They always measure performance and productivity from the

customer’s perspective. For a call center, managers will usually measure the

percentage of customers whose issues are addressed in a single call.

Among the metrics JPF found it needed to change was processing time. The company

had traditionally measured the time from when an application arrived at the newbusiness

processing center to when the approved policy was printed and bound. JPF

switched to measuring the gap between when the application is mailed to the company

and when the adviser receives a completed policy, which is how its customers assess

the company’s speed. Customer-focused metrics of this kind helped erode the “My

work is all that matters” mind-set of JPF’s employees.

Another important principle of lean production is that shop -floor goals should be linked

to the metrics that are applied to the CEO’s performance. Toyota calls this hoshin

kanri , or policy deployment, and it is the best way to align an organization’s activities

with its strategic objectives. At JPF, the metric for the CEO’s performance is the ratio

of the company ’s total acquisition expenses to the value of new paid premiums. The

cell’s productivity directly affects this measure—as productivity increases, the

acquisition expense eventually decreases. An employee inputting applications is

evaluated by the number of applications he or she inputs per hour, and the input

team’s manager is assessed on the hourly number of applications the team inputs. The

input manager’s boss, the New Business assistant vice president, is assessed according

to the productivity of the input team and all the other steps in the process. These

productivity rates affect the metric for the VP in charge of New Business, a measure

that is the same as the performance metric for the CEO. Thus, the CEO’s success is

directly linked to each frontline worker’s productivity. In this way, JPF has spread

accountability and rewards throughout the system, rather than concentrating them at

the top.

Companies like JPF whose production systems rely heavily on third-party vendors also

need to look at their suppliers through a lean -production lens. What JPF found

prompted it to establish new vendor-selection criteria such as alignment with Jefferson

Pilot objectives, aggressive annual goal-setting, and the adoption of lean processes

that fit well with JPF’s. It replaced one of its vendors with a company that not only

provided faster turnaround times at a lower cost but also was willing to commit to

ongoing performance improvements.

JPF was luckier than some companies: Most of its existing frontline metrics matched

lean -production requirements fairly well, and for those metrics, the team immediately

established short-term (three-month) and long-term (12-month) goals. In areas where

metrics needed adjustment, such as productivity by job function and certain customersatisfaction

...