Work monitored electronically: ten guidelines that make the process more effective and humane - Employee/Employer Rights

Ten guidelines that make the process more effective and humane.

"The daily television rental will be $3.47 with tax. Would you like that added to your room bill?" "Yes, thank you," replies Walter Davis. Naomi Ruten, services rental agent, enters the automatic billing code into the hospital's ancillary services screen. As she hits the Enter key, four things happen. First, Memorial Hospital's Total Patient Care (TPC) system adds the first day's rental charge to Walter Davis' billing detail. It then sends an electronic message to the housekeeping subsystem (HKP), telling the staff that Walter Davis ordered a television at 9:30 a.m. that should be delivered as soon as possible to Room S214. Third, TPC adds the first day's charges to Naomi's daily rental income total. Finally, it increases the count of transactions Naomi has processed and logs the amount of time she spent handling Davis' call.

Darren Simcoe reads the order on HKP, removes a television from inventory, and tells the system that unit #1073 is being delivered to Room S214.

While Simcoe connects the television, Dr. Lenore Green checks in on Davis. She uses a terminal keypad next to Davis' bed to send a medication order to the hospital pharmacy. She also sends a message to the senior ward nurse to measure Davis' vital signs at 3:00 p.m. TPC transmits the order and message to the appropriate recipients. It also logs that Dr. Green was with a patient in Room S214 at 10:10 a.m.

Simcoe enters his employee ID number into the keypad, along with the transaction code to indicate that the television has been installed in the room and is in working order. HKP subsequently transmits a copy of this information to TPC.

Simcoe enters Room N387, where he removes an oscilloscope that is no longer required. Before leaving, he logs his ID number and the activity on the keypad in the room. Again, TPC receives a copy of the transaction information, along with its entry time.

At Memorial Hospital the total Patient Care system and its subsystems process thousands of such activities daily Most information it handles has a direct bearing on patient care or accommodations, such as the order for medication and television billing. Other items, like the inventory unit number for Davis' television, are vital to the hospital's administrative control or facilities management. As a by product of the computer-mediated operations, TPC also processes an enormous amount of data about the staff and its performance.

The practice of accumulating this information is called automated performance measurement, computerized productivity monitoring or work monitoring. Critics call it electronic surveillance, eavesdropping and a tool of the electronic sweatshop. "This system uses computer programs and software to record, direct or evaluate the work employees perform, on a computer."

Monitoring work electronically is not new. It has been used in manufacturing and assembly for years to track throughput, inventory and scrap. Labor costing systems have long been used to gather information to allocate costs for accounting purposes and provide the necessary data to set piece-rate wages or bonuses. What is new is the use of similar systems in services and among professionals.

Memorial Hospital's system, for example, can tell management that: * It took Naomi an average of four minutes and 15 seconds to complete a rental contract. * Dr. Green visited four cardiac and three surgery patients between 10:00 and 11:30 a.m. * Simcoe's activities are unaccounted for between 10:15 am. and one hour later when he removed the oscilloscope from Room N387.

All of this information can help hospital managers measure productivity, control costs and schedule staff needs. It can provide useful feedback to employees about their performance and help them improve their work habits. It can also create information overload for managers who receive more data than they can realistically absorb. More important, it can increase stress for employees who fined themselves feeling that their every move is watched by the electronic eye of the silicon supervisor.

Electronic work monitoring has been a topic of debate for more than five years. Labor unions argue that the systems are demeaning and stressful; some employees say that they find monitoring by impartial silicon supervisors preferable to evaluation by subjective humans.

Monitor design and acceptance

Dr. Chris Higgins and I conducted one of the largest studies of computerized performance monitoring carried out to date. To gather information about the acceptance of monitoring and its effect on attitudes toward work, we analyzed surveys from 1,500 employees of 51 Canadian service firms. The firms represented 14 industries and subindustries, including telephone, publishing, airlines, investment, banks and trust companies, insurance, and postal and courier services. The study included both women and men.

Both monitored and unmonitored employees replied to our survey, and we compared their attitudes. Without such comparisons, it would have been impossible to attribute monitored employees' attitudes and behaviors strictly to the presence of a monitor. Among monitored employees we included people who worked under a variety of systems.

As Exhibit 1 shows, employers used computers to track many different aspects of performance.

We used the employees' reports of how performance was measured, rather than the employers' descriptions of the evaluation systems. Our own research has shown that employees may or may not correctly understand how their performance is measured. However, it is not how an employer actually measures performance that matters, as much as what an employee thinks is measured.

Very few firms used any software to full potential, or even to the degree of the Memorial Hospital example. The majority used computers to track the number of transactions, error rates and/or keystroke speed. By including various monitor designs, we were able to look at the effects of a range of monitoring approaches, from unobtrusive and low-level counters to extremely invasive surveillance.

Ten guidelines for managers

The study provided the following 10 practical guidelines for managers to use monitors effectively and humanely:

1. Choose tasks deliberately. Some behaviors, such as being courteous to customers, simply do not lend themselves to quantitative measurement or numeric surrogates (i.e., the number of customer complaints is a surrogate for courtesy). Employees find monitors most tolerable if they are directed at tasks employees feel are important job tasks, and if those tasks can be accurately reflected in quantitative terms.

2. Consider recipients carefully. Many companies post individual performance data on bulletin boards, in the belief that seeing their colleagues' numbers will motivate less-productive employees and reward higher producers. Are your employees motivated by competition? In some companies, internal competition is effective and healthy. But in others, broadcasting individual performance can be extremely disheartening to employees. It can trigger resentment of high producers and generally raise the level of stress and anxiety.

Think about performance data from the need-to-know perspective. If the system is meant to provide intrinsic motivation and rewards to the employees, consider a system in which only the employee sees his or her individual data. If rewards are not based on the measures collected by the machine, think about whether anyone other than the employee and an immediate supervisor or manager needs to see individual measures. If the system is used for extrinsic rewards, it still may not be necessary to broadcast results. The wider the audience for the data, the less acceptable monitoring systems are to employees and the more likely they are to have a negative effect on attitudes and actions.

3. Monitor regularly. Most service work is input-driven. A salesclerk can only sell to customers who are in the store, a telephone operator can only respond to incoming calls, and the workload in an emergency room is higher on Saturday night than Wednesday afternoon. These factors can all affect response time, error rates and productivity. Similarly, there are peaks and valleys in an individual's performance, even with a steady stream of work. Some individuals work better early in the day, others at midafternoon. One day an employee will have a headache, the next day someone else has a pending parent-teacher conference nagging in the back of his mind. Unless it is very carefully designed, intermittent monitoring may not reflect the effects of an uneven workload.

Service workers have discussed intermittent data collection with us in interviews and through written comments on the surveys. Their comments demonstrated the difficulty of proving to employees that any form of intermittent monitoring accurately samples the peaks and valleys in workload. In the face of all evidence and arguments to the contrary, many employees continued to believe that intermittent monitoring caught all of their slack periods and considerably fewer of their peaks.

4. Correctly attribute performance. A monitor design should attribute responsibility for performance accurately. Interdependent tasks, such as processing a transaction through many work centers, are often the wrong targets for individual performance measurements. Errors or time delays may not show up at the work center in which they occur, but only become obvious later in processing. The person handling the problem may inappropriately be held responsible for the problem. A more productive employee may be seen as slower than his or her colleagues.

Many managers or control system designers design monitors to automatically gather and report data about every individual. This is not necessarily desirable. If your employees have highly independent tasks, and bonuses or wages are tied to individual performance, you must accurately capture the work of the individual; group averages are not enough. But if you are trying to instill a sense of teamwork, gathering individual statistics alone can be self-defeating. Aggregates of group performance are needed. In fact, you may not want to keep any worker-specific data at all.

5. Supplement your monitor with data on work quality. It is easy to print out transaction counts, average response times, slack times and error counts. It is not as easy to describe or evaluate less tangible aspects of performance. Yet, it is vital that you have a system for measuring performance and that it gets regular, visible use. Employees should be told about their qualitative performance as often as they hear or read about their production.

The message you give your employees about the importance of production and service influences their attitudes. The design of your monitor is a significant, but small part of that message. You can convey your message through the way you collect information about qualitative work, how you handle customer complaints and special requests and what you say during training and promotion. These and other factors interact with the design of the monitor to create an overall perception of what you and your company value.

6. Stress nonquantitative factors equally. Many companies that monitor or use manual methods to gather the same information often believe they give equal weight to intangible and unmeasurable aspects of performance. But when one looks at their evaluation systems, it is not clear that the intention to consider other factors is ever translated into action. Supervisors who keep comprehensive files of quantitative performance data may have nothing more than an occasional scribbled note about any other aspect of an employee's work. Scribbled notes may be triggered only by bad performance and may not reflect exceptionally good work. Employees of such companies quickly get the message that it is their production that is remembered and rewarded.

7. Provide a mechanism for employee participation and error correction. For monitoring to be acceptable to employees, they must believe it is accurate, appropriate and complete. Employees can reason or argue with a human supervisor when they do not feel the supervisor's information represents their performance fairly. They can point out factual mistakes in the data or lobby to have other factors considered in the evaluation. One cannot carry on such a discussion with a monitor, and many companies have no mechanism for correcting or attaching explanations to data. Employees may not see the data until well after they have been collected and, in some companies, employees never see their own performance information.

In our survey, and in other studies that have explored this issue, the participants felt strongly that they should be involved with the silicon supervisors. This does not mean that employees must always be actively involved in choosing or designing the monitor. Instead, they need to know that the measures incorporated into the machine represent what they would consider fair, accurate, appropriate and complete information about their work. They should also have the opportunity to review information collected about them and to challenge it effectively.

8. Train your supervisors. We have heard many complaints from service employees whose supervisors overuse monitor data or let the monitor take over the supervisory task. There have been so many complaints, in fact, that it is tempting to think there's no such thing as bad monitoring, only bad management. However, there clearly are bad ways to monitor. While it is easy for supervisors to rely heavily on quantitative data from a monitor, particularly when performances are weak or inadequate, supervisors need to be trained in the role that monitor data should play in staff evaluations. This is true for any form of quantitative performance data. It is especially important for information collected by silicon supervisors.

9. Demonstrate the role of the monitor explicitly. Employee interpretations of how the monitor figures into their performance evaluations continue to surprise managers, supervisors and researchers. During one series of interviews, 25 claims processors described their understanding of the reward, bonus and promotion criteria in their company. Their descriptions of what was monitored, how important the data was to bonuses and promotions, and how the data were collected varied dramatically from one person to another. An observer asked if all of the processors worked for the same company. Imagine her shock when told that they not only worked for the same company, they all had the same manager

Employees will form their perceptions of the monitor through its apparent daily use. This makes it especially important that companies not confuse employees through the actions of supervisors and managers. If the data is meant for staff planning, supervisors should never refer to an employee's data in a performance evaluation. If monitored activities are only one component of overall quality performance, other components must be given equal weight. For example, when an employee who scores highly on the monitored elements is promoted, the manager also must clearly describe the individual's strengths in other areas. Your employees probably have a good idea where they each rate in terms of response time, sales volume or output. It is less likely that they know how they compare to one another in terms of leadership potential, creativity or dealing with difficult customers.

10. Keep electronic monitoring in perspective. Above all, it is vital that you keep the data and the functions of the monitor in context. Quantitative data and silicon supervision should be but one part of a complete evaluation and control system. In some cases, it should play no part at all. Properly designed computer programs will always count, add and average more accurately and consistently than human beings ever could. But human supervision, observation and listening to employees and customers about the pace and quality of work will always provide information that cannot come from software. insightful managers know that this information will continue to play a major role in any service environment.

Rebecca Grant, Ph.D., is an assistant professor of information systems at the University of Cincinnati in Ohio. She is the author of Silicon Supervisors: Computer Monitoring in the Service Sector.

The author gratefully acknowledges the support of Labour Canada and the University of Western Ontario in funding this research. The University of Cincinnati subsequently supported the preparation of this article.

COPYRIGHT 1992 Society for Human Resource Management
COPYRIGHT 2004 Gale Group