Modeling Dynamics of gaining expertise in a call center
Abstract
Knowledge and expertise are the most precious assets of a call canter, which enables the staff
to serve the customers on high quality. As call centers have a tense working condition, staff
are likely to leave this job to a better position in the company or outside, so a real challenge
for the managers of these centers is first to moderate the working condition , and then to cope
with the high rate of turnovers. This article aims to the last, by modeling the cycle of training
a new staff and building experiences, which helps him to serve the customers. By entering the
risk of leaving job, this model is objected to find scenarios to reduce the costs of turnover.
Keywords: call centers, knowledge Management, expertise
1. Introduction
One of the most important elements of CRM (Customer Relationship Management)
concept is call-center, as customer contact point (or touch point). (Fickel, 1999)Usually IT-
related technologies (such as fax, telephone, Email or automatic dispatching and response
systems) are used to provide appropriate response ( or support) for the customers. CRM aims
to provide this service by integration of all other support and contact channels.
In most CRM models, KM (Knowledge Management) has been pinpointed as one of the
most important critical success factors in the long term.(Eckerson and Watson, 2000) Infact
KM tools and strategies could be used to enrich the CRM practices in most cases. Many
efforts have been concentrated on implementation of KM principles in CRAm, especially
Call Centers.( rasooli and albadvi,2007) but as these centers are really various, approaches
have been diversified. In this article, author has concentrated on KM view of a call center,
which includes the experience and knowledge building for staff during working period in call
center. This has been done by building a dynamic model, to capture the current situation of a
real call center, which is aimed to support the customers of a software company in Iran. This
model has been used to test some of desired solutions, and then has been used to propose a
better policy.
2. Literature Review
In this session, two concepts are examined more carefully. One is about the nature of
knowledge and experience, and the other is the nature of call centers, discussing the necessity
of knowledge management.
2.1. Knowledge and Experience Management
Although there is no agreement about the meaning of knowledge, during years of knowledge
management appearance, delegates of different philosophies, has added their own definition
of knowledge to this field. Knowledge as a structure or an atom is defined as “‘systematical or
intelligent understanding that is used for doing effective actions in direction of system
objectives” and in contrast definitions such as “justified correct idea” (Nonaka and
Takeuchi,1995) and “complete usage of data and information together with skills,
eligibilities, beliefs, cognitions and motives of organization’s manpower’( Kalseth, 1999),
have used other points of view for defining knowledge. Without any doubt, there is a close
relation between experiences and knowledge. For instance we can consider experiences as
what refines knowledge, or as a special (purified) kind of knowledge. Knowledge and
experience both are considered as spiritual assets (Sun and Finnie,2005), although they are
treated differently.
Davenport and Prusak define knowledge as: “Knowledge is a fluid mix of framed experience,
values, contextual information and expert insight that provides a framework for evaluating
and incorporating new experiences and information.” (Davenport et al, 2000) According to
them, Experience is a part of knowledge that is in close relation with the function of human's
mind.
Although each kind of research in context of knowledge without considering experiences
appears to be meaningless, giving exact definition of experience is extremely difficult.
Generally, experience can be defined as previous knowledge or as skills one can attain in his
or her own daily life.( Bergmann and Schaaf,2003) Experience is previous knowledge which
is obtained by solving faced problems for which successful solution has been provided.
Knowledge processing is only a fundamental prerequisite for solving problems and having
experience is also necessary for repelling actual difficulties. Storing knowledge is a necessary
condition for storing experiences. Knowledge and experience both are abstractions at
different level. Experience is more abstract than knowledge, because it is often in the form of
meta-knowledge. Transferring people’s experiences and turning them into knowledge for
other people was always an important problem. Storage and analysis of knowledge can be
reduced in creating new experiences (Bergmann and Schaaf,2003).
2.2.Call centers and their functionality
Most organization are using call centers as the main channel for interaction with their
customers. This centers have many benefits such as increasing efficiency, increasing
working/serving hours, minimizing costs while increasing flexibility which caused expansion
of business. The main challenge of these centers is to provide timely and accurate
information(Robertson,2002)
Call center association (Call Center Association,2008) defines call centers as "managed
physical or virtual operations in organization which a group of people are mostly working
with telephone, fax or computer. Sometimes staffs have not special skills, which have to
prepare suitable response in a short period. However most call centers use high skilled
workers which work under special service level agreements (SLA). (Rasooli and
Albadvi,2007)
Robertson ( Robertson,2002 ) believes that high turnover and tensed working environment
are the most challenging factors for call centers. The turnover is about 26% for full time
personnel and about 33% for part time staff.
In most CRM models, KM has been involved in order of decreasing cost and improving the
quality of service for customer. (Demerest, 1997). In fact for developing a reliable
relationship with customer, support should be customized by having the knowledge of his
own needs. Most CRM methods not only mention the transactions’ data, but also stores these
data in databases and repositories, so they could be used to harvest knowledge. CRM
processes are almost mentioned as Knowledge intensive processes ( Eppler et al., 1999)
which knowledge flow management (- to the customer, and also —from the customer) is
really crucial.
In call centers, it may happen that one staff is serving customer, or customers are using self
service support. On each situation, some knowledge users should have access to the
appropriate data , information and knowledge to provide instant correct and consistent
answers. Not only collecting, quality issues and structuring efforts are necessary for
converting information to the knowledge, but also human expertise is a vital element for
making it usable. Therefore in call centers, staffs have a crucial role, and the role of IT tools
for supporting their knowledge and expertise can not be ignored.
This article focuses on the human expertise, and building it through the working time of a
call center staff as an instant.
3. Model
This model is built to demonstrate the state of Staff's experience during working period in
the company. Model is started by hiring staff, and ends when he quits, which would take
about 3 years. the assumption is that he'd quit when the anxiety is more than his threshold.
The main part of the model is a state variable which demonstrates the expertise of a sample
support staff( who is called the "agent" in the model). This variable shows a mix of
experience, data and information which are consistent with the davenport's definition of
knowledge. This pool could be filled either by formal training programs provided by the
company (the rate variable named "training" ) or by serving the customer, which is consistent
with this fact that experiences are generated by real problem solving, which in this case refers
to customer service function. Sometimes company hires more professional guys (educated in
the related academic fields) which makes the initial value of this stock to be higher, otherwise
it is supposed to be 0 when a new staff is recruited. There is also a decreasing rate variable
which is forgetting, which works always as a decreasing factor for the stock of knowledge,
and its effects are prominent when the agent does not work as service provider (for instance
on a working break)
When a new staff is recruited, he should be trained in order to get ready to serve to the
customers. These trainings are demonstrated as a learning package in the model, as it is done
in the company for the new staff. Other trainings are also provided during the work, as
needed. The need is identified by comparing the agent expertise with the needed expertise (a
balancing loop)
The expertise of the agent defines the level of service that he could provide for the customers.
Basically there are 3 levels of expertise for the technical staff, which are assumed to be
gained by a technician in 3 years. These levels are set to the variable named "desired quality
of service" which indicates the expectations of the company. This makes the staff need to
work and get updated, either by serving more challenging customers (learning by doing and
learning from their peers during the job) or taking training courses offered by the training
department of the company.
There is another stock variable assumed in the model as customers. It shows the customer
assigned to the staff. This should be increased during the working age of the employee. If this
number increases, the tension of employee increases, and he may leave the company. By
serving them better, they got more satisfied, and bringing more customers for serving ( the
assignment of customers to the staff is done by a line manager, who balances the workload as
well as assuming quality of service) as staff serve more customer, they have more chance of
learning by doing , so they become more expert.
The whole model could be seen in figurel.
time to accomlish
training time to forget
training
o:
Agent Expertise
forgetting
need for training
arming by serving
ality of service customer
“ Oo
service
axdety treshold__,,
‘decide to quit
Fe capatity of service
stocked
andety
increase in amdety
Figure 1- structure of model
4, Simulation and scenario testing
Some tests show that the model is consistent with the reality. First of all, Model has been
built and tested partially. This included building the expertise stock and it's related rates first,
and then adding the training loop (making all the other rates as constant variables set to zero)
the second phase was making forgetting rate work. Then adding the customer assignment to
work, makes the learning rate active. And at last, quitting the job has been added by
considering the anxiety to the working condition. The results show that:
- The expertise of the agent is increasing gradually, at first the main reason is training,
then it is completed by learning by doing, which is a result of increasing the quality of
his work.
Figure2- simulation of model- current situation
Current
Agent Expertise
80
40
0 9 18 36
Time (Month)
- Although training is considered to have positive effects, this model shows that the
decision to quit is just occurs after a training ( in month 34) it is consistent with the
observations in the field. Because training usually is followed by the more customer
assignment which results in increasing the anxiety.
- Quality of service decreases especially after yearl, which logically should be the
prosperous era of agent's work. This is mainly because of the anxiety, which is cause
by assigning new customers to the staff. In the real situation this is known as the main
reason of agents’ quit.
By assuming that the model could regenerate the real trends, the boundary is set to be
optimal, and the scenario testing started. The main solutions that seemed to be appropriate
were:
- Training customers in order to make less errors, and then they'll have less need to
support. This decreases the pressure on the agents by each customer (increases the
capacity of each agent to serve) this may increase the working period of the agents
(increases the quit time from 34 to 38). As it may seen in the model, this is just a
temporary solution that does not change the patterns of variables. The problem that
still forces line manager to assign more customers to the agents, still remains.
Figure 3- scenario of training customers
Cureet]
Carent
Agent Exoertse
8
‘Tooe (Mout)
- Another solution that is proposed id using IT capabilities to provide solutions for the
customer. This may include a wide range of solutions, from automatic (machine)
responses to the providing a FAQ for the support staff. This is shown in the model by
decreasing the training need of staff , decreasing the chance of learning by doing
(providing standard solutions from a prepared repository) and as a result, the capacity
of serving could be increased. This means that staff only learn by training courses
(which occur on a regularly basis of 8 month) and the anxiety increases more slowly,
so the staff remains more in the company (from 34 to 55 months)
5. Conclusion
By simulation of model, decision makers ( CEO and support manager) found out that 2 of
their main solutions have only short term effects on the tendency of personnel to leave. So
they became eager to find another solution for the long terms. By working with the model, it
seems that there are 2 main points that could be considered for building the solution. One
point is that the anxiety variable has no decreasing flow, which could work as a moderator for
the staff. At the other hand, the root of support problem are the software (product) bugs ,
which is touched by the support staff but they could not do anything about it. This resulted to
establish a feedback between the software development team and the support team. This
means that the software bugs are reported to the R&D team. This changed the model, as it
could be seen in figure 3.
ume to accomusn
training time to forget
a
Agent Expertise
forgetting
an for training reported bugs} software bugs
learning by serv — | oe
ldecresing bugs
quality of service customer
new Oo iiian |
anxiety
decide to quit
_ of service
f ==
Gereue Wane increase in anxiety
—ee
axiety treshold_,
Figure 3- modified model by considering a feedback from support to the development
(R&D) team.
In the modified model, a new stock has been added. This stock is related to the software bugs,
which is a index to show the quality of software. Some times it does not mean that software
crashes, it means that some modifications should be done to be suitable and customized for
the customers. This is done by reporting bugs to the R&D team. This takes time for
implementation, which is embedded by 4 months delay in the model. Also a decreasing flow
has been added to the stocked anxiety model. As the reporting loop has been build, it could
act as a empowerment factor for the support team which makes personnel more relaxed about
their work.
As this model runs, we could see that agent expertise does not increase , and is steady at a
certain level which is satisfactory for the customers. Also there is just one course of training
needed ( after hiring personnel) which cuts company costs, because these technical training
are done by experienced staff of company. The other important factor is that the decision to
quit is not taken in 5 years period (which by now is a target for the company) , and it would
happen in month 70, which is far beyond the expectation.
Current3
Current
Agent Expetise
80
40
Mf |
0 5 3) 45 60
Time Month)
Figure 4- simulation of model with R&D feedback activated
The examination of other important variables ( stocked anxiety, quality of service and need
for training) shows that these variables have more desired patterns. Anxiety increases more
slowly in the modified model, which makes the less decrease of quality of support. Also
personnel could learn from their day to day experiences, which makes them feel their
working environment as a learning atmosphere, and decreases their stress.
6. References
Bergmann:R.;Schaaf«M.;2003; On the Relations between Structural Case-Based
Reasoning and Ontology-based Knowledge Management; Proceedings German
Workshop on Experience Management «April «2003 - informatik.rwth-aachen.de
Bergmann:R.;Schaaf«M.;2003;Structural Case-Based Reasoning and Ontology-Based
Knowledge Management: A Perfect Match?; Journal of Universal Computer Science «
vol. 9 «no. 7 «608-626
Call Center Association;2008; 60 Ideas in 60 Minutes:2008 Session; available from
http://www.nationalcallcenters.org/reports.php
Davenport , Thomas; De long; david; Beers ; Micheal; 2000; successful knowledge
management projects; in knowledge management yearbook 1999-2000; By James W.
Cortada, John A. Woods; Butterworth-Heinemann.
Demerest,Marc; 1997; Understanding knowledge management; Long Range
Planning; Volume 30, Issue 3, June 1997, Pages 321-322, 374-384
Eckerson, W. and Watson, H. (2000), Harnessing Customer Information for Strategic
Advantage: Technical Challenges and Business Solutions, special report, The Data
WarehousingInstitute, Chatsworth, CA.
Eppler, M., Seifried, P. and Ro” pnack, A. (1999), ‘‘Improving knowledge intensive
processes through an enterprise knowledge medium’’, Proceedings of the 1999 ACM
SIGCPR Conference on Computer Personnel Research, ACM Press, New Orleans, pp.
222-30.
Fickel, L. (1999), “Know your customer”, CIO Magazine, Vol. 12 No. 21, pp. 62-72.
Kalseth, karl (1999) knowledge management from a business strategy perspective ;
Journal of FID review. vol.1, No, 37-41.
Nonaka ,I., Takeuchi,H.; 1995; The Knowledge-Creating Company: How Japanese
Companies Create the Dynamics of Innovation;
Rasooli,p.,Albadvia.,2007;knowkedge management in call centers;electronic journal
of knowledge management;vol 5, no 3,323-332.
Robertson ,James; 2002; available from Step Two Designs website on
www.steptwo.com.au; last accessed January 2010.
- SunsZ.; Finnie «G.;2005; Experience Management In Knowledge Management;
Faculty of Commerce — Papers (University of Wollongong); available from
http://ro.uow.edu.au/commpapers/33
