Performance analysis – Final Report
1. Introduction – the project in a nutshell
(The issues are real, but many names are pseudonyms for privacy sake)
The scenario is global, since the issue is related to the offshore-outsourcing of jobs. VTrack is a
vendor organization in India, which caters to the companies in the USA, by setting up off-shore
outsourced office units in India, including the hiring of manpower.
In July 2004, VTrack set up a call center for STE, a telecom company based in the US, in
the silicon city of India. Now the company is losing customers in the USA, as well as receiving a
large number of complaints via e-mails and postal mails, on poor service quality when customers
call the customer service lines. Back in India, there is a high turnover too. There are 350
workstations at the call center.
Couple of years back, in 2002, STE had outsourced information data processing units and
account processing units to India through VTrack. Those units had no complaints and are
functioning smoothly till date.
When I visited India in May, my friend Geeta, who is an associate of VTrack, wanted to
know a whole lot about call center functioning in the US. At that time, all I knew was, when we
called for service, we got the required information or assistance. The service is quite fast and
efficient. I could not help Geeta much at that time. Nevertheless, whenever I saw any media
coverage ( and we had plenty as part of election campaign!), I paid attention to call center news.
Later, when I read about the performance analysis project as part of the IT7150course, I
talked to Geeta. So, basically, Geeta acts as my proxy in India to close the intercontinental gap of
physical distance that I am faced with.
1
2. Goal(s) analysis
The goals are dichotomous.
The primary goal of this performance analysis is to remove the dissatisfaction of the STE customers
who approach the customer call lines for service.
The secondary goal is to reduce the high turnover of the call center agents hired.
3. Purpose of the analysis
The purpose of the analysis is to identify the glitch (es) that generates customer dissatisfaction and
also the high turnover and of course, to bring in intervention and to improve the performance.
Apparently, the root cause of customer dissatisfaction as well as the call center agent dissatisfaction
(generating the turnover) is to be tracked, quantified and removed.
Customer
dissatisfaction High turnover
Impact on
revenue,
profits
4. The target population analyzed
Since the job analyzed is a direct interactive task – telephonic customer service - the target
population analyzed fall into two categories. At the receiving end are the customers (the affected
2
population) residing in the United States, who have subscribed to STE Telecom. On the other
end, there are the call center workers (also called as agents), who are stationed in India.
5. Accomplishment and the performer
The performer is identified as the call center agent.
The accomplishment in this scenario is telephonically providing accurate information to customers
for their queries or problems regarding STE operations.
At this point, I went ahead and tracked the specific operations addressed by the call center agents,
which fell into five categories –
o billing,
o address change,
o new subscription,
o technical support,
o general information
(Later, all categories were discreetly taken care of while interviews were held.)
6. The methods of data collection
Because of the intercontinental spread, data collection posed as a big challenge.
I was not able to physically visit or observe the call center agents at task. So that left me with the
other two vertices of the triangle – document analysis and the interview.
A. Document analysis
Again, since I was not officially in contract with STE or VTrack, I had to ask for specific details I
needed through Geeta and had no direct access to records.
The documents I requested for and scrutinized were mainly statistical.
3
1. Percentage of customer complaints prior to outsourcing
2. Percentage of complaints after outsourcing
3. Customer growth graph
4. Turnover
B .Surveys
I requested to conduct a survey .Surveys were mailed to 50 customers (with the postage pre-paid),
spread over the United States.
The survey questions were few, but direct –
1. Are you satisfied with the STE customer call service lines and the agents?
(Yes/No)
2. If „No‟, what is the main problem you encounter –
a. Discourtesy?
b. Delay?
c. Technical ignorance?
d. Clarity?
3. How do you rate STE customer call service when compared with other utility customer call
service providers (like the electric power company) you deal with?
a. Better
b. Same level
c. Worse
C. Interviews
Since India does not have strict laws needing any prior approval before interviews, this was the
easiest tool I could use. I requested Geeta to schedule telephonic interviews with at least 20 out of
the 350 agents and two other administrators of VTrack.
My interviewee list included
1. Call Center agents in India ( 20 –> at least 2 from each of the five main categories –
billing, address change, new subscription, technical support, general information)
2. VTrack administrators – 2
3. If possible, 5 STE customers in the US.
4
Since interviews were going to be my major data collection tool, I used both Gilbert‟s Behavior
Engineering Model and Mager and Pipe‟s analysis model to frame the questions.
Gilbert‟s model was used to detect the environmental supports, while interviewing the VTrack
administrators.
The reason I used Mager & Pipe was because, according to Geeta, VTrack was blaming attitudinal
problems and laziness on the part of the agents, because they (the agents) were now earning mega-
bucks they could never dream of. I told Geeta that we had a scientific and systematic way of
analyzing the issue by addressing all possible causes.
D. Call monitoring” -
I could not think of a way to identify the exemplary performer through interviews alone.
Performance had to be measured somehow. Then I remembered my mobile telephone company‟s
call monitoring permission message: [“Your call may be monitored for quality purposes……”]
When I checked with Geeta, I came to know that though there was official call monitoring,
no analysis is done with the recordings. I drafted some measures and gave them..
1. Number of interruptions or prompts like “Excuse me”, “Say that again”, “Pardon” from
both ends
2. Number of pauses – from the agent‟s side
3. Duration of each call
4. Number of calls ending cordially – ( Proof of customer satisfaction)
A measure I am still awaiting is the average duration of a call before outsourcing.
7. Results
A. Document analysis
There was a 2% decline in customer growth in the past five months.
The turnover was 52% in eight months.
The company receives on an average 13 customer dissatisfaction notes daily.
5
B. Survey, interviews
I failed to interview the five STE customers. As is common practice, before my self-introduction
was over, they hung up. It was no surprise to me, since I often do the same, when I get a similar
call.
Luckily, maybe to give vent to the disappointment and fury, STE received back 33 surveys that
were mailed. The best part was, though my questions were objective, there were additional
subjective comments from the customers, which, to a large extent, compensated for my failure in
conducting interviews.
Interview report – VTrack administrators
Reinforcing what I had heard from Geeta, all they had to say was how the high salary had changed
the attitude of the hired agents, who were showing lapses in duty, laziness etc.
Interview report -Call Center agents
This was the most productive part of my analysis techniques, I should admit. The agents were very
articulate. My inferences were:
o Majority of the agents were very young, fresh from college
o Their main attraction is the high „foreign rated‟ salary
o The facilities were rated as okay by some, but as „suffocating‟ workstations by a few
o The other comments were „boring‟, „same job‟
o Two of the comments made me reflect on other avenues to investigate. The comments were:
1. “We are nothing but the „electronic housekeepers‟ (referring to „monotony‟).
2. “We work on „graveyard shifts” (referring to „night shifts‟).
C. Results – call monitoring measures
o Only 22 out of the 383 calls monitored in a week‟s time ended on a cordial note. That is less
than 10%.
o Of the above, 4 were directed to the same workstation (no. 187) and 2 to workstation no. 73.
These results did not leave us with enough scope to identify an exemplary performer, but 187 was
marked as a comparatively better performer.
o The duration of the call was too long with three to four prompts and pauses at least.
6
8. Other avenues investigated
As I had mentioned, the disappointing statements by the call center agents and the measure of the
long duration of the calls left me thinking for a couple of days.
The survey reports had also shown „delay‟ and „clarity‟ as the issues.
To me, everything converged to a mismatch of the required „skills‟ to accomplish the task.
I requested to interview the „hiring officials‟ of VTrack on the criteria used.
9. Findings
1. The main criterion for hiring was the transcript and the educational qualification. So, the
hired agents are often technocrats, unemployed, in a developing, poor nation like India.
2. The personal interview and record check lasted for less than twenty minutes.
3. The candidates were told about night shifts, but no „biological clock‟ orientation was done in
this regard.
4. All agents talked to, were satisfied with the job-aids provided by STE at the workstation.
At this point, I went ahead and filled up Gilbert‟s Behavior Engineering Model.
Data Instruments Incentives
√ OK √
Knowledge Capacity Motives
√ X ?
The problem zones
1. Instruments : HVAC – poor, electric grid – unreliable, data/voice backup, breakdowns
7
2. Capacity : lack of skills for a VOICE job – „accent‟, delivery ( speed, clarity), sleep apnea
3. Motives : lack of intrinsic motivation - no career growth for the high qualification
Media reports and the connection to VTrack hiring procedure
I did research on media reports that I had come across on cross-cultural outsourcing constraints. In
particular, I was reminded of the 11alive news feature by Jim Lehrer on the New York Times
correspondent‟s visit to outsourced organizations in India and the „accent neutralization‟ classes.
I checked with VTrack if they had used this accent neutralization class for training. The VTrack
response was that since they had catered to STE before while outsourcing data processing jobs with
no training, this time too, they had just acquainted the hired agents with the workstation by a day of
orientation.
10. The final investigation
Thanks to the performance analysis class, I knew there was still a gap left unaddressed which can be
measured only by comparison with similar „cross-cultural‟, jobs.
I contacted another „head-hunter‟ service agency in India, to find out what other jobs had the
highest market in the USA other than information technology. I was informed that „nursing
professionals‟ were recruited in large numbers.
It is a chance coincidence that, to my luck, nurses too needed the „voice job‟ skills, since
they have to essentially converse and communicate orally with the patients on a regular basis.
On further investigation, I found out that the recruiting criteria were much stricter. Apart
from a passing score on CGFNS, the subject knowledge test on nursing, a high percentage score in
IELTS (International English Language Testing System) – both written and „spoken‟ English tests
rated separately, was a criterion for selection.
That reminded me of TOEFL scores in addition to SAT or GRE, for foreign applicants for
admission to universities in the USA.
8
11. Recommended interventions
(i) To acquire and test the required skills
o All hired agents need immediate „voice job‟ skills training. They need to attend the
„Accent neutralization classes‟, since the US and Indian accents have great disparity.
o A passing score in competent tests like IELTS „speaking‟ test is mandatory for
„voice jobs‟ like call center workstation jobs. Keep a deadline of six months, that is,
by June 2005, for all existing employees to pass the test. This may also be
considered as part of the „change management‟ process.
(A similar example may be cited in Georgia – all GA teachers – old and new - need to
pass the InTech test before June 2006, now that technology is an integral part of
education – „changing with time‟. Otherwise they will lose the certification).
(ii). Time zone difference orientation
o Just by informing the candidates of the time zone and night shifts, they are only
mentally oriented. This is an act that requires literal „physical‟ orientation which
directly affects the efficiency. So, every hired candidate must be given other
interesting, light tasks for a week, during the shift timings they are scheduled for
and thus be oriented physically.
(iii). To reduce the turnover
Only 1/5th of the services addressed are of the technical helpdesk. Hence recruitment criterion
should not be based on technical qualification alone. This would reduce the issue of the lack of
intrinsic motivation. Candidates with arts and humanities background will derive better intrinsic
motivation from such communication and interactive jobs and will stay longer, reducing the
turnover. Another alternative is to hire from a higher age-group, than the early twenties.
(iv) Better infrastructure
Power supply is to be improved and the interior of the call center office is to be re-designed to
minimize the congested feeling in the workstations.
9
12. Data based conclusions
STE had saved from 20% to 50% on operating costs, when it implemented off-shore outsourcing of
the technological data processing and programming jobs to India a couple of years ago. The prime
reason was the low wage structure in a developing nation like India. On the other hand, India has
rich manpower resources by way of highly qualified technical hands. So the job quality was well
maintained too.
Poor cultural affinity (the cultural gap not identified or predicted)
Both STE and VTrack had overlooked the difference between the technical data processing task and
the skilled voice job to be performed in customer call centers.
Now, the poor cultural affinity between offshore agents and American customers has risked the
customer satisfaction and retention problems.
52 % of Turnover – an irony in competencies and behaviors
o Most of the Indian agents now hired are college educated, professionally qualified,
ambitious, in early 20‟s, lured by high salaries to the call centers.
o The irony is that the competencies and behaviors that made the Indian technocrats to
perform successfully in the outsourced IT jobs are the same competencies that are
hindering the call centers. The dynamic agents are craving for intrinsic motivation,
making them hunt for greener pastures.
Time mismatch – Biological clock and time zone difference
o (Graveyard shifts) – When it is 2.15 a.m. in India, it is 3.45 p.m. in Atlanta the day before
and 12.45 p.m. in Los Angeles.
13.The immediate action plan and the continuing issues
Since it is a question of losing the customers, interventions have to be effected on war-footing.
o Accent neutralization class is the priority. Even after that training, the style of talking may
have cultural disparities. So it is not a permanent solution.
o The language skills test must be mandatory with a deadline (like InTech for teachers).After
the tests, filtering and firing of some of the existing employees is inevitable.
10
o The effective way for summative evaluation of the interventions is to continue with the
quantified measuring standards set up for „call monitoring‟ while we analyzed the problem.
o After calculating the training and intervention costs, it would be wise for the management of
STE and VTrack to hold an audit to check if the anticipated 20-50% profits in outsourcing
are still met with. So should such voice skill jobs be outsourced after all?
o The current hiring procedure is the reason behind turnover. So, the recruitment criteria
should be changed. Overqualified professionals will move away soon if there is no intrinsic
motivation or career growth. As such, call center workstation job is more of an articulate
task.
o There is definitely going to be an initial „quality‟ issue between the natural accent and style
of the US resident customers and the delivery of the trained accents of the Indian locals. Of
course, history states that Demosthenes, the famous orator, fought against his birth-defect of
stammering and remedied it by hiding himself in a cave and practicing to articulate by
putting pebbles in the mouth and speaking! Thus there should be no compromise on rigorous
quality training.
- The end of report –
11
Appendices
12