Management Summary
Kodak maintains two support centers across the nation. These centers receive calls from hospitals and clinics
worldwide. The call center in Dallas is made up of two levels. Level One takes calls directly from the customer and attempts to solve the problems through standard
troubleshooting. The purpose of this first level is to filter out the simple problems from the more complicated ones. Less skilled and lower paid technicians handle this in order to take the load off of the higher skilled and higher paid technicians in Level Two and in turn lowering Kodak’s support costs.
One problem that we felt was relevant was that employees were underestimating their actual applied work time. Another problem that we felt was evident was
teamwork within a case. Only the person who closes the call in Vantive, the computer call tracking system, gets credit for the call. Also, discussions between colleagues about work related problems or cases are never acknowledged or recorded. Additionally, there were reported problems with dialing into the necessary equipment to fix the customer’s problem.
As calls come in, employees are responsible for answering the calls in a timely manner, documenting the
customer’s problem, as well as steps they took to fix it, and recording their applied time for the call. As it stands, Kodak has estimated an approximate productivity rate for which they would like their workers to reach; this rate stands at 70%. When originally evaluating the applied time data, Level Two employees were only credited for
working 39% of the time. We wanted to figure out if this was an accurate portrayal of what the call center employees were doing, or if this was a false representation.
After gathering data, we concluded that two main elements were missing from the estimated productivity calculation. For instance, not only should holiday and sick leave be taken into consideration, but a buffer for poorly estimated applied time should also be included to create a more accurate estimate. After reconfiguring the productivity calculation, we feel we have reached a
realistic estimate, which also yields an acceptable productivity rate.
Background
Eastman Kodak has been involved in health imaging for longer than a century. Today Kodak is deeply involved in the capturing, processing, viewing, printing and storing of images for diagnostic and non-diagnostic applications. For obvious reasons, hospitals are Kodak’s primary customer for the health imaging division.
Hospitals rely on Kodak Health Imaging products for a variety of patients, from the ten-year old getting an
annual check-up, to the trauma patient who needs immediate attention. For this reason, Kodak must maintain a twenty four hour a day call support center to deal with the
problems that occur with their machines. These problems can range from mechanical to computer related and
everything in between. In the case of Kodak Health Imaging products, Murphy’s Law prevails, “If anything can go wrong, it will,” and never when you expect it.
Kodak maintains two support centers across the nation. These centers receive calls from hospitals and clinics
around the world. The call center in Dallas is made up of two levels of service. Level One takes calls directly from the customer and attempts to solve the problems through standard troubleshooting. The purpose of this first level is to filter out the simple problems from the more
complicated ones. Less skilled and lower paid technicians handle these calls in order to lower Kodak’s support costs by taking some of the load off of the higher skilled
technicians in Level Two.
One problem which we felt was relevant was that employees were underestimating their actual applied work time. Another problem we felt was evident was crediting teamwork within a call. The calls are tracked on a system called Vantive. In this system, only the person who closes the call gets credit for it. Therefore, if a call
technician put in time on the call but then transferred it to a colleague, the person who started the call would not receive any credit. Also, discussions between colleagues about work related problems or cases are never acknowledged or recorded. Obviously, if two colleagues are working on a problem together, they are both still working, and should therefore both be given credit. Unfortunately, their collaborative time cannot be tracked in Vantive. In addition to these problems, there were also reported
difficulties with dialing into the equipment necessary to fix the customer’s problem. This also added additional time to the case, but went unreported.
Analysis
Kodak had already estimated the productivity rate of their typical Level Two employee. This consists of a 52 week period, in which a typical full time employee works 40 hours per week. Holiday, sick leave, vacation, training, meeting, and checking email were all taken into
consideration for the target productivity rate. As calls come in, employees are responsible for answering the calls in a timely manner, documenting the customer’s problem, as well as the steps they took to fix it, and to record their applied time for the call. As it stands, Kodak has estimated an approximate productivity rate for which they would like their workers to reach; this rate stands at 70%. When evaluating the applied time data, Level Two employees were only credited for working 39% of the time. We wanted to figure out what was happening the other 31% of the time.
As there are no preset timers, it is up to each employee to give an honest estimate of the time they
worked. As it is easy to lose track of time while you are working, one of our main goals was to see if the call
technicians recorded accurate estimates. To go about checking these estimates, we designed a data sheet to record the events of each call. The first piece of
information to gather was the actual amount of applied
time; this included dialing in to the servers, going to the lab, consulting with fellow employees, reassuring the
client, as well as all other activities that went into resolving the customer’s problem. Also, we recorded the employee’s actions, like whether or not they looked at the call history or if they had any trouble dialing into the system. We also recorded whether or not they were the last employee to close the call, and whether or not the problem was solved. Each of these factors helped us to see where employees were spending their time. A copy of the
collection data sheet that we used may be viewed in Appendix A.
This method of data collection assumes that the data collector is accurate with time keeping and maintains awareness of the situation that the employee is dealing with. This method also assumes that employees did not change their behavior due to our presence. Although we tried to make them feel very comfortable and told them we would not be recording their names or any case numbers, there is still a good chance that employees skewed their data or acted differently due to our presence.
Technical Description of the Model
After reviewing the data, seen in Appendix B, we found the average underestimation of applied time to be 2.167 minutes. From this data, we were also able to resolve that Level Two employees handle approximately 3.36 cases per day, with an average of 2 calls per case. Therefore, at 3.36 cases per day with 2 calls per case, each being
underestimated by approximately 2.167 minutes; this results in about 56 hours per year for each employee of unreported applied time on a call.
Other findings, lead us to believe employees do not perceive consulting with fellow colleagues as working. This does not include chattering near the water cooler, but merely discussing cases and many times, helping a fellow employee with a call. As we observed the employees for at most four hours at a time, we can contest that each
employee easily spent around 20 minutes assisting a
colleagues at various times. As they did not have a case open or a call number for the help they were providing, it always went unreported. Therefore, we included ‘colleague assistance’ in our model, for which we estimated 45 minutes per day for each employee. This resulted in 173 hours of unreported applied time per year for each Level Two
Another estimated error in reporting time includes reviewing call history but not including it in applied time estimates. This was very hard to track as it usually took them anywhere from a couple seconds to 2 minutes; however, most employees seemed only to track their time from the moment they picked up the phone, not from when they opened the call history.
The last problem that we came across was simple to fix with a slight adjustment of its estimate. Call technicians were originally credited as spending one hour per day on voicemail and email. After observation, we felt that this estimate was a little shy of what it should be. Using the same process to come up with a valid estimate that we used to find unreported time, we deduced that the actual time would be around one hour and ten minutes per day.
Our model is a simple algorithm. Based in a Microsoft Excel spreadsheet, we calculated a revised estimate of
applied time. This spreadsheet may be viewed in Appendix C. This algorithm looks at the annual estimated work hours and takes training, vacation, sick leave, holiday,
underestimated time, colleague assistance, checking email, and meeting attendance into account to calculate the
Analysis and Managerial Interpretation
At 40 hours per week for 52 weeks a year, the annual work hours estimate is 2,080 hours. Looking at Appendix C, you can see each breakdown of time. Indirect work hours totaled to 921 hours. These were subtracted from the total number of annual hours available to arrive at an estimate for potential direct hours; this number ended up to be 1,159 hours per year for each full time equivalent employee. According to Vantive, Kodak’s productivity
tracking system, an average, full time, Level Two employee reportedly worked 738 hours a year. By dividing the actual annual direct hours by our new estimate for potential
direct hours, we get a 64% productivity rate. This
improves the rate of productivity by 25%. Kodak Level Two employees are only missing their mark by about 6%.
Therefore, by taking the additional ‘unreported time,’
‘colleague assistance,’ and the new time spent on voicemail and email into consideration, Kodak has a more realistic evaluation of their employees.
Conclusion and Critique
From our observations, we can firmly conclude that two elements were left out of the original productivity
time’ and ‘colleague assistance.’ ‘Unreported time’ consists of time employees do not report due to
underestimation. ‘Colleague assistance’ is exactly what it sounds like: the time spent while helping fellow co-workers on cases. In addition to these added estimates, we also developed a more realistic estimate of time spent on
voicemail and email. By including the two new elements and our revised estimate for email and voicemail in the
Productivity Forecasting Model, Kodak’s Level Two
employees’ productivity rate increased to 64%. This yields a 25% increase.
As we feel like this estimate is solid and valid, we also acknowledge the fact that it is based solely on our observations in the Call Center. Human error should be counted as a factor when taking this data into
consideration. For further implementations on our study, we would suggest asking Level Two employees where they feel they spend most of their time and if they feel they are being credited for all aspects of their job. As our observations are based over a two month period, it is highly possible that there are other duties for the Level Two employees which we did not get an opportunity to