Data Collection Methods

In most cases, a company's total direct labor cost is not so high that it warrants the creation of an elaborate data collection system. Instead, you can either focus on a simple system that collects only the most basic data or else install a system that utilizes a greater degree of automated data collection, thereby keeping costs low while still obtaining a high degree of detailed information.

If a simple data collection system is needed, the easiest possible system to implement is one in which employees are assumed to work 40 hours per week and the only need to log hours is to record any overtime, which is recorded on an exception basis and forwarded to the payroll staff; the staff then enters the additional overtime costs into the payroll system and generates payments to employees. This approach is most useful when a company has a relatively fixed base of direct labor employees who rarely work any additional hours and who also rarely work less than a fixed number of hours per week. A further justification for such a system is when a company has such a small amount of direct labor cost that a more elaborate timekeeping system would not be worth the effort to implement. This system yields no information whatsoever regarding how the cost of labor is being charged to various jobs. It has the singular benefit of being very inexpensive to maintain, but at the cost of providing no costing information to management.

A slightly more complex system is to have direct labor employees fill out time cards that itemize their hours worked each week. These time cards are reviewed by their supervisors for accuracy and then forwarded to the payroll staff, to compile the information and keypunch it into the payroll system. This approach is most useful when there is a significant amount of variation in the number of hours worked per week, resulting in continuing variations in employee pay from week to week. This approach requires considerably more administrative labor because of the large amount of data entry involved; additional labor is also needed to verify the entries made by employees and to investigate and correct any errors.

One step up from this entirely manual system is the addition of a time clock. In its simplest form, a $100 to $500 time clock can be mounted on a wall, into which employees can insert their time cards to have their "in" and "out" times recorded. This approach makes time cards easier to read and controls the recording of time worked, so that there is less chance of any deliberate alteration of time worked. This approach is highly recommended, since the additional cost is minimal and is easily justified by the increased level of data accuracy.

The next step up in system complexity involves the use of a computerized time clock. Like the time clock just described, this device is also mounted on a wall for employee access, but it contains two additional features. One is the use of a bar-coded or magnetically coded employee card that is "swiped" through a channel on the side of the clock whenever an employee clocks in or out. This card contains the employee's identifying number; the system records that number, and all associated time worked, with complete accuracy. The second innovation in the clock is a computer that is linked to a central payroll computer. This feedback mechanism allows the time clock to reject employee swipes if they are made at the wrong time (such as during the wrong break time) or are made for employees who are not supposed to be working during specified shifts (which may occur if one employee brings in someone else's card and attempts to record that person as being on the premises and, therefore, eligible to be paid). The system can reject swipes that fall into any number of violation parameters and require the override password of a supervisor to record those swipes. The benefits of this innovation are twofold: it yields a great improvement in a company's control over the timekeeping process, and all data swiped into it requires no further keypunching—all of the data is sent straight into the payroll system, where it is reviewed for errors and then used to pay employees. This eliminates the cost of extra data entry, as well as the risk of data entry errors. These innovations do come at a price, however, which is typically in the range of $2,000 to $3,000 per automated time clock. A large facility may require a number of these clocks if many employees must use them, so the cost of this addition must be carefully weighed against the benefits.

A larger volume of data can be obtained by using the just-described computerized time clocks at every workstation in the production area, or a modified version thereof. By doing so, employees can easily punch in information about which jobs they are working on at any given time, without having to walk to a centralized data entry station to do so. These workstations can be time clocks that are directly linked to the payroll system; but since these clocks are so expensive, this option is not normally used, especially if many workstations are required. A more common approach is to purchase a number of "dumb" terminals, which have no internal error-checking capacity at all, and link them to a central computer that does all the error checking for employee and job codes, as well as hours worked. This option is much less expensive, especially for very large facilities. However, it suffers from one significant flaw: If the central computer goes down, then the entire system is nonfunctional; this problem does not arise when using automated time clocks, for each one is a separately functioning unit that does not depend on the availability of a central computer. This problem is a particular issue in companies that have large amounts of machinery that generate electrical energy, for the extra radiation can interfere with the transmission of signals from the workstations to the central computer, usually either requiring the installation of heavily shielded cabling or the use of fiber optics, both of which are expensive options.

An employee uses the dumb terminal to enter his or her employee number, then the start time, and then the job number. All time accrued from that point forward will be charged to the entered job number, until the employee enters a different job number. This data entry process may require a large number of entries per day, which introduces the risk of a high degree of data inaccuracy. The problem can, however, be minimized by the use of bar-coded or magnetic-stripe employee cards, as previously described, as well as bar-code scanning of all current job numbers.

This last option is clearly much more expensive than any preceding option, since the cost of the central computer can be anywhere in the range of $10,000 to $250,000, and requires a large number of dumb terminals that cost at least $500 each. What is the reason for incurring this expense? This system enables a company to track the time worked on specific jobs. This is a very important capability when customers are charged based on the specific number of hours that employees work on their projects, especially when the customer has a right to investigate the underlying hourly records and to protest billings that do not match these detailed records. This is a particularly important issue for government work, where cost-plus contracts are still common, and the government has a right to closely review all supporting labor records. It may also be a major concern for any organization for which the cost of direct labor is still a relatively large proportion of total costs; otherwise, managers would have no valid information about how a large proportion of company costs are being incurred. Nonetheless, the data entry system required to support the collection of this information is very expensive, so you should conduct a cost-benefit analysis to see if the value of the supporting information is worth the cost of the system.

It is also possible to have employees manually track the time they charge to each job on which they are working. Though this option may seem much less expensive than the use of the data entry terminals described, this approach is not recommended unless the number of employees using it is very small. The reason is that the level of data errors will be extremely high, given the large number of jobs to which labor is charged each day (the time charged to a job may be wrong, as well as the job number to which the time is charged). As a result, the cost of the administration time required to track down and correct these problems will greatly exceed the cost of installing an automated time-tracking system; this correction cost will be so high for a large facility that the comparable cost of an automated system will be far lower.

A final timekeeping system to mention, one that is not frequently used, involves backflushing. In fact, this is not a real timekeeping system at all; instead, the standard labor hours are stored in the labor routings database for each product and multiplied by the amount of production completed each day, which yields a standard amount of labor that should have been completed for each workstation in order to complete the total amount of production issued. This method is only good for developing approximations of the amount of labor that was needed to complete each step in the production process. It is of no use for spotting labor inefficiencies and cannot be used to derive payroll (since it does not report hours worked at the employee level, nor would these numbers be accurate even if it did so). Thus, the backflushing method, though a simple way to derive approximate labor hours, does not yield accurate information for most purposes to which direct labor information is applied.

It should be apparent from this discussion that a higher degree of data accuracy and a lower cost of timekeeping on a per-transaction basis can only be achieved with a high degree of expensive automation—and the more information required from the system, the more expensive it will be to collect it. Accordingly, you must first determine how badly a company needs each possible type of direct labor data, then structure the data collection system based on the level of need. Before making this decision, it is best to review the following section, which describes the various types of data that can be collected through a timekeeping system.

Super Business - Project Management Articles

Data Collection Methods

Data Collection Methods