�

Balancing Service Costand Service Levels

The Service Op�miza�onChallenge White Paper

Copyright Notice Copyright © 2011 ClickSoftware Technologies Ltd. All rights reserved.

No part of this publication may be copied without the express written permission of: ClickSoftware Technologies Ltd.

Publication Notice The information contained herein does not constitute a warranty of any kind. ClickSoftware Technologies Ltd. reservesthe right to revise this publication and make changes without notification. ClickSoftware Technologies Ltd. assumes noliability for losses incurred as a result of information contained herein.

�

�������������� White Paper . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .3

The Service Optimization Challenge . . . . . . . . . . . . . . . . . . . . . . . . . .3

Executive Overview . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .3

Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .4

Avoiding Losses from the Start . . . . . . . . . . . . . . . . . . . . . . . . . . . . .7

Pre-defined Capacity vs. Dynamic Appointment Optimization . . . . . . . .7

Dividing Your Territories is Not Enough . . . . . . . . . . . . . . . . . . . . . . . .8

Grid-based vs. True Travel-based Optimization . . . . . . . . . . . . . . . . . . .8

Reducing Travel Time . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .9

Linear Distance-based vs. Street-Level Routing-based Optimization . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .9

Are You Really Balancing the Load? . . . . . . . . . . . . . . . . . . . . . . . . .10

Grid-based vs. Dynamic Load Optimization . . . . . . . . . . . . . . . . . . . .10

Managing the Inevitable Change . . . . . . . . . . . . . . . . . . . . . . . . . . .11

Daily Batch vs. Continuous Background Optimization . . . . . . . . . . . . .11

Staying Competitive . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .12

Code and Core Changes vs. Configured and Component Changes . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .12

Summary . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .13

�

����� ��� �� ����One of the greatest challenges in service optimization is

increasing customer service levels while reducing operational

costs. It is often difficult – if not impossible – to achieve one

without sacrificing the other.

The reason is the inherent conflict between sending the most

cost-effective resource and sending the best qualified one.

Service operations must also take into account a number of

other factors when dispatching field resources, including

geography, parts required, breaks, unforeseen emergencies,

service levels, and other constraints. A relatively small workload

of 100 service calls in a single eight hour day makes it virtually

impossible to effectively balance all of the simultaneous

calculations and decisions required to ensure happy customers

at the lowest possible cost.

Some scheduling systems with less evolved optimization

methods solve this problem by narrowing down the possible

choices in a sequential step by step approach, where each step

attempts to identify the best technician and/or time, based on

a single rule. The result is that later in the process, only those

technicians remaining as candidates from the previous step can

be considered. When one considers even a simple example of

overtime versus geographic location, it is obvious that this

approach often results in sub-optimal schedules.

The financial penalties can be significant. For example, a

service operation with 300 field resources that loses only 15

minutes per day, per resource (for excess travel, idle time,

repeat visits, etc.), wastes more than 18,750 hours of

productivity each year. At an hourly cost of $50, it totals

$937,500 – excluding overtime wages and mileage costs.

When you consider all of the costs that a service organization

incurs – including lost productivity in the field workforce, labor

spent managing and dispatching the schedule, and the cost of

missed Service Level Agreement (SLA) penalties or lost

customers due to missed commitments – it becomes easy to

see how a small percentage of inefficiency can create

significant costs.

This white paper discusses some of the factors that contribute

to operational inefficiencies in field service. Traditional

approaches to solve these optimization challenges are

examined, as well as their impact on the service “balance

sheet”. Starting with making arrival commitments to customers,

through creating, and maintaining a schedule – we identify some

of the key challenges dispatchers face each day.

This paper also provides an overview of how scheduling

optimization software applications can significantly contribute to

efforts to eliminate these problems. It examines optimization

features such as dynamic appointment scheduling, travel-based

optimization, street-level routing, dynamic load balancing, and

continuous background optimization. Finally, we discuss ways to

avoid costly application code and core changes associated with

implementing a new enterprise application.

Whilst this White Paper focuses on the operational scheduling

process, where the decision horizon is typically 7 hours to 7

days, this is only one link in the service decision-making chain.

As we move away from the day of service, the decision making

horizon stretches from 7 weeks to 7 months in advance for

tactical resource planning that typically involves vacation and

training planning, and 7 months and beyond for strategic

capacity planning and forecasting. The rewards for managing all

these in a synchronized manner are enormous

Using the analogy of the slot machine, you only hit the jackpot

if you have all the ‘7’s’ on the win line. The holistic ‘777’

approach to service management looks at the service operation

at 7 months, 7 weeks, and 7 days prior to service, and

synchronizes strategic, tactical, and operational decisions. With

good forecasting,

your resource capacity planning in the various regions will be at

just about the “right levels”, making your daily schedule as

“uneventful” as a daily schedule could possibly be.

Future White Papers will explore in much more detail the other

links in the service chain including forecasting, capacity long-

term planning and tactical resource planning.

�

�����������Although you may not be familiar with the “Service Optimization

challenge”, the bottom-line of your service operation probably

suffers from it every day. Optimizing service simply means

scheduling resources in the most productive and cost effective

way to ensure sufficient customer service levels. The challenge

arises out of the complexity of finding the best “overall” field

engineer for every job, every time. Not just the closest, or the

one that can respond fastest, but also the one that best

balances the workload, minimizes overtime, and has the parts

needed to complete the job the first time.

Some service operations may claim they meet this challenge

well because their dispatchers “have been doing this for years”.

There is little doubt that an experienced dispatcher – given

enough time – can usually make an optimal decision for having

the right

person, at the right place, at the right time. But within this

statement lies the challenge. As service workloads increase,

two things happen:

(1) The amount of time allowed for each scheduling decision

decreases

(2) The number of options that should be considered for ‘who’

can do ‘what’ increases

The following two charts demonstrate why no field workforce is

really immune:

Chart #1 shows a comparison of a call volume (25-300 calls

scheduled per dispatcher) vs. the amount of time (within an 8

hour day) allowed to make each decision. At 150 calls per day,

there are less than four minutes allowed for each decision.

300

250

200

150

100

50

0

25.0

20.0

15.0

10.0

5.0

0.01 2 3

Calls Per Scheduler Per Day Minutes For Each Decision

4 5 6 7 8 9 10

Chart #1At 150 calls per day, there are less than four minutes allowed for each decision

�

Chart #2 demonstrates the mathematical complexity of making

the optimal decision, as the number of options increases. With

only a few more options (e.g. a different order of completing a

day’s route), the complexity grows well beyond the capacity of

even the fastest, and most experienced dispatcher.

10,000,000

1,000,000

100,000

10,000

1,000

100

10

1

1 2 3

Number of Decision Points Number of Op�ons to consider

4 5 6 7 8 9 10

Chart #2With only a few more options (e.g. a different order of completing a day’s route),

the complexity grows well beyond the capacity of even the fastest, and most experienced dispatcher.

6

In order to overcome these challenges - service companies have

traditionally made compromises in optimization that allow easier

decisions. A simple example is the creation of ‘invisible’

boundaries of coverage, i.e. any service call received within a

certain territory will be dispatched to one of the technicians

servicing the area.

This does reduce the number of possible choices and the

complexity of the decisions for dispatchers, but it has a cost.

For example this disregards the overall workload balance in the

field and the possibility of scheduling a resource from a

neighboring territory that is actually closer to the customer. The

underlying assumption that accompanies this is that “losing”

5% or 10% of potential optimization is tolerable – and maybe

unavoidable.

Considering that a 10% increase in productivity for a field

workforce of 300 resources can be worth millions of dollars

each year, competition makes any inefficiency intolerable. In

this paper, we will provide practical insight into how

organizations can overcome the Service Optimization Challenge.

Technicians CustomersFo

r Who

m

Where

Who

WhenW

ith W

hat

What

Parts Time

GeographyTasks

The service optimization challenge arises out of the complexity of finding the best “overall” field engineer for every job, every time.Not just the closest, or the one that can respond fastest, but the one that best balances the workload, minimizes overtime, andcan arrive within the committed timeframe. This requires considering numerous variables, in addition to the technician, customer,

and specific task – this includes geography and travel time, parts pick up and inventory, service level agreements and much more.

7

������� ����������������������� ����������� �� � � ���� ������������������������������� ���������

������������ �����!������

Most service scheduling applications – and manual processes –

provide customers with commitments for arrival based on a

predetermined number of slots for each geographic area,

product type, and time window. These slots are allotted to

customers arbitrarily as they call, until they are all filled.

The problem with this approach is that it does not take into

consideration a number of factors, such as the location of

previous service calls, or the true availability of individual field

resources, including their skills. This can lead to problems with

service efficiency, effectiveness and utilization; for example, the

excess travel that results from customers who are located on

the same street, but who arbitrarily select appointment times

that are hours apart. When appointments do not consider the

varying duration of different types of work, late arrivals and idle

time result from differing job durations.

When appointment booking is optimized, service organizations

are able to identify criteria that are important to them - such as

shortest route between calls – and take these criteria into

account to determine a time that is still convenient for the

customer, but also maximizes workforce efficiency. Optimization

looks at the existing workload versus workforce capacity in

terms of geography, skills and time, rather than a predefined

estimate. The schedule can also be continually re-optimized

throughout the day, creating further improvements in resource

utilization while maintaining customer commitments.

Simplis�c

Morning A�ernoon

4 miles

Intelligent Clustering

Shortest route - 36 miles Shortest route - 24 miles

In the first diagram, the scheduler simply allotted slots according to customer preferences without considering the location ofexisting commitments. In the second, a service optimization application automatically suggested appointments that would

complete the same service calls, along with a 33 percent improvement in travel mileage.

8

Benefits as a result of optimized appointment scheduling:

• Improved utilization by reducing the amount of time spent

driving

• Increased customer satisfaction by providing reliable

appointments

• Reduced time required keeping customer commitments

!� ���� �"�������������#��#�$������ �%&���'(�#��� #)�������� ��'(�#��������������Some service operations attempt to reduce unnecessary travel

by dividing their territories into zones. Although this is a step in

the right direction, it can also lead to inefficiencies for

customers near the zone borders. Although another resource

may be nearby, he will not be assigned because it is over the

line or boundary.

Assigning resources to fixed zones, regardless of the size of the

zone, will always result in some loss of efficiency. Traditionally,

human schedulers handled these situations manually; however,

as call volumes increase, it becomes impossible to constantly

identify and resolve these situations. In addition, many

automatic scheduling applications actually exacerbate the

problem by requiring resources to be associated with a zone.

Service optimization solutions use travel distance between

calls, rather than pre-assigned grids, to look for the closest

resource to complete a call. The result is that the resource with

the minimum travel time to reach a call is consistently assigned,

while resources generally remain in their own area.

Subsequently, the optimization of travel time is significantly

improved, regardless of geographic area or distribution of calls.

These two diagrams demonstrate two resources in bordering

territories that are required to travel an additional 16 miles to

service four customers. Consider only 100 customers serviced

each day, and more than $100,000 may be wasted each year.

*���+��#��#�����#����+���� ��'(�#��������������%• Reduced travel time to service the same number of

customers

• Reduced travel mileage costs to service the same

number of customers

• Increased workload capacity from the existing field

workforce

2 Miles10 Miles10 Miles

2 Miles

9

,���� ���� �������%�-������!�#����'(�#��� #)�������'-� ��,���� '(�#����������������Some scheduling applications now consider travel time using

the linear distance between service calls, based on postal

codes or pre-defined travel times between service areas.

However, this is a more granular method of defining territories,

and suffers the same inefficiencies for customers near postal

code boundaries. Furthermore, postal codes may cover a metro

area that requires 15 minutes to cross, or a rural area covering

100 square miles.

While this approach works well in principal, linear distance

solutions are not able to take into account important details

that can have a large impact on arrival times. This can create a

disaster on a call-by-call basis when one-way streets, bridges,

bodies of water, or other obstacles stand between one customer

and another.

The results of linear travel estimations are not just excessive

travel costs, but also unhappy customers that expected a field

resource to arrive on time. Even if each technician makes only

one minor mistake per day, the resulting customer discontent

can be extreme.

Advanced service optimization solutions prevent these kinds of

routing mistakes from happening. They include street-level

routing features that incorporate detailed GIS (Geographic

Information System) data. They take into account obstacles

such as bodies of water, bridges, one-way streets, parks,

campuses, and posted speed. In addition to reducing travel

time, this method of optimization ensures that customers are

given commitments you can keep.

When this service call was assigned using linear travel, it did

not take into account that the red support technician would have

to travel to the next bridge and back—resulting in more than

twice the travel time of the green support technician. Street-

level routing would have prevented the waste of four miles, 10-

15 minutes, and the associated costs.

Benefits as a result of street-level routing:

• Reduced number of missed calls and late arrivals

due to inaccurate travel times

• Less time and money lost related to travel,

by scheduling the actual closest resource

• Fast identification of closest resources for emergency calls

using a map display of the schedule

Customer

10

.���"��,����/�*������ �����-���0%�&���'(�#��� #)�!/�����-���������������Balancing the workload among resources presents a significant

challenge to many service operations. Often, the number of

service calls per technician is based on average call distribution

– but distribution varies from day to day. When this happens, the

problem is felt throughout the organization, from service

technicians who are constantly under or over-utilized, to upset

customers who wait for service appointments that aren’t kept.

The same applications that use grid-based scheduling to

minimize travel time typically use the same grid-based method

of load balancing – with the same inconsistent results, and the

same consequences. A one percent annual loss in utilization for

a workforce of 300 will cost more than $350,000. Losing the

loyalty of customers will cost, in the long run, even more.

By dynamically balancing the workload without the constraints of

grids, service optimization solutions can always schedule based

on the true and current workload for each resource. Similar to

the manner in which they consider real travel time, these

solutions consider each call to be assigned along with the best

resource, based on their workload.

The first diagram demonstrates how giving one resource more

work than can be handled in one day leaves another under-

utilized. The second shows how dynamic load balancing results

in 100 percent utilization and happy customers.

Benefits of dynamic load balancing:

• Increased utilization of resources in the overall schedule

• Increased number of completed service calls per day

• Increased employee satisfaction as a result of not being

overloaded or under-utilized

11

1��� �� �������� ���(������� �%�!���/�*���� #)��������#�*�2 �����������������How does a service operation manage their schedule in light of

the unexpected events that occur every day? Even the best

person or application can only consider the information available

at the time of schedule preparation (e.g. the night before

service). When changes occur on the day of service, as they

inevitably do, a schedule that was initially optimized now suffers

from work gaps due to cancellations, or missed customer

commitments as resources get delayed in traffic or onsite.

To lessen the impact of ‘same-day’ changes, most scheduling

applications optimize on a periodic basis, when users are not

making changes to the schedule. This is due to the heavy

performance toll that optimization can take on the application,

along with the complexity of optimizing a schedule that may be

changing simultaneously.

What is needed is an advanced architecture and techniques that

allow periodic dynamic schedule optimization, according to

application-specific processes and policies, throughout the day.

This enables the entire schedule to be continuously optimized –

resolving conflicts and increasing efficiency - without impacting

the dispatching process. This capability takes into account

newly received calls and other unexpected events and

reschedules calls to ensure that resources do not return to the

same area twice unless absolutely necessary (e.g. to meet two

emergency SLAs).

This simple example of three separate service calls shows how

a lack of optimization resulted in a missed commitment for one

resource, and only 50 percent utilization for the other. With a

workload of 300 field engineers the annual impact of losing only

one percent of utilization across the entire schedule could be

approximately $350,000. Benefits as a result of continuous

optimization:

Benefits of dynamic load balancing:

• Increased overall utilization of resources

• Completion of more service calls per day

• Reduced need for manual or supervisory intervention to

adjust the schedule

4 Hours (#1) 8 Hours (#3)

4 Hours (#2)

Op�miza�on

UnfulfilledCommitments

50% U�lized

MadeCommitments

100% U�lized

4 Hours (#2) 4 Hours (#2)

8 Hours (#3)

12

���/�� ���������� �%������������������� �#� #)����+� ��������������������� �#�Finally, companies that want to cut costs while improving levels

of service to their customers have some difficult decisions to

make. Some of those decisions include:

• Whether to invest time and money to build a solution, or

make significant code changes to an existing application

• How to best address changes – both planned and

unplanned – that will take place in the business during the

time it may take to deploy an optimization solution

• How much development will be required, along with the

additional cost and support necessary to adapt the solution

to business changes

• How to accommodate changes in scheduling policies and/or

market conditions as quickly and cost-effectively as possible

Traditionally, the core logic of scheduling applications has been

inaccessible to service organizations, requiring the customer to

make changes to their policies and processes, or the vendor to

make changes to the scheduling logic. These changes can be so

painful to a company’s competitiveness that the organization is

forced to choose between the cost associated with an endless

development cycle and the cost of not changing at all.

One answer to this dilemma is a graphical user interface (GUI)

for configuring the logic of the scheduling system – including

which scheduling policies are considered, and how each one

behaves.

If an organization has varying scheduling policies (e.g. for

dif ferent territories, product lines or businesses), easy

configuration through such a tool allows the application of

specialized scheduling policies to only specific calls and

resources

This figure shows a GUI display for adjusting a load balancing

policy within an application. All scheduling policies can be

configured through similar windows and the majority of

scheduling adjustments can be applied to an operational

system with no interruption to the scheduling system.

13

�����/�As these examples have shown, the costs associated with

human scheduling or a rudimentary scheduling application can

quickly add up. Choosing the appropriate scheduling

optimization solution can enable your organization to achieve

significant advantages in many areas. These include:

• Significantly reduced travel time required

for completing calls

• Improved levels of customer service through reliable and

responsive schedules

• Increased utilization of your resources

• Reduced time to manage the schedule

• Flexibility to handle dynamic changes within your

business both short and long term

As a final note we would like to provide insight into the broader

picture of service optimization. There are many other factors

that influence the quality and business value of service

workforce management. For example, if you don't have enough

service engineers for handling tomorrow's load, you're faced

with some tough choices and “firefighting”. Putting in place a

solution that includes monitoring, data analysis, forecasting and

planning could make the organization run smoother and more

productively. However, this is out of the scope of this white

paper.

.(������2��+�����ClickSoftware® is the leading provider of automated workforce

management and optimization solutions for every size of service

business. Our portfolio of solutions, available on demand and

on premise, create business value through higher levels of

productivity, customer satisfaction and operational efficiency.

Our patented concept of 'continuous planning and scheduling'

incorporates customer demand forecasting, long and short term

capacity planning, shift planning, real-time scheduling, mobility

and location-based services, as well as on-going communication

with the consumer on the expected arrival time of the service

resource.

As the pioneers of the 'W6®' concept more than 20 years ago,

we have perfected solutions for solving a wide variety of

problems on Who does What, for Whom, with What, Where and

When. The combination of proven technology with educational

services helps businesses find the right balance between

reducing costs, increasing customer satisfaction, employee

preferences and industry regulations/legislation.

ClickSoftware's solutions manage over 200,000 resources in

service businesses across a variety of industries and

geographies. Our flexible deployment approach, breadth and

depth of solutions and strong partnerships with leading

CRM/ERP vendors and system integrators makes us the

number one choice to deliver superb business performance to

any organization. The company is headquartered in the United

States and Israel, with offices across Europe, and Asia Pacific.

For more information, please visit www.clicksoftware.com.

Follow us on Twitter.

��� " #�$��%&#'����(���)����*��������+�������������!�� �������33*���� ���4�1.�3�53���� 65557���5���354�685�7��8���93�,�- 685�7��8��:�39

&.# �&���(���)�������;����/������,���*�����4�*2#�-��8-<��� =����637�:�5�:38333,�- =����637�:�5�:3833�

���(���)����������&/�����0��$>������-���#��)��9��*:3����?���2+������1���!��#�������� =��9�637�:9��595���3,�- =��9�637�:9��595���99

��'����','���(���)��������������� ����.������;��24������*����� 9����'>���#�� ���,���;��*����:98;����'��2 ���9��8�#������� =�98����8:��9�33,�- =�98����8:��9�3�

���(���)�����/ ���������1�����-� ������:�@����������1��(����4�A���333��� =:��637��99�:�:�33,�- =:��637��99�:�:�3�

���(���)����2�����3�3���?��������������4��:'���2�� ��2�'��4����(/�'2��2/�4�B������� =5�������:��::�,�- =5�������:�����

���(���)����'��������A���2��.��������?����4�*��2'*������C���4�&��+����#��,���&� ����C�����33�>��/���4��������� =9������������55

�������'����������?����������+����������(������2��+�����

�#������)��2#�+�����)�����������#���#D��2#�+�����)��