The classical view of inventory data collection is that of employees filling out forms of various kinds throughout the warehouse and production areas, which are then forwarded to a central data entry location, where hordes of clerks keypunch the data into a central computer database. Although this was a reasonably accurate view of the situation in the past, the types of systems available for collecting information are now more efficient and effective. These systems were developed because of a growing recognition that traditional data collection methods require a great deal of employee time that could be better spent on value-added tasks. Also, having a secondary data entry step increases the likelihood of keypunching errors, which can be completely avoided by some of the data collection methods discussed in this post.
Some of the data systems that can be used to collect inventory information are shown in below graph. They lie along a continuum that begins with loosely formatted data, such as that found on a faxed document, and ends with perfectly formatted data that can be directly entered into a computer system without alteration, such as electronic data interchange (EDI) transactions or transactions entered through an electronic form. A special case is document imaging, which can be tightly coupled to a company’s computer systems or maintained as a freestanding system with no linkages at all. Accordingly, it is surrounded by a larger box in the exhibit, indicating the range within the exhibit that it can occupy. Based on the information in the exhibit, it is evident that an inventory accountant should recommend installation of the systems noted in the upper right-hand corner because they provide the best means for collecting the highest-quality costing information that can be injected directly into a company’s central database of costing information.
Let us say that Omega Company wants to track the progress of a product through every step of its production process. Being a technologically advanced organization, it has installed data entry keypads at each of its workstations. One of these products is assigned the part number AD-546-798. The operator of each workstation is required to enter the part number using a keypad, followed by the number of units completed. The inventory accountant uses this information to determine the progress of work-in-process batches as they move through the plant. However, the part number is so meaningless that 3 of the 10 workstation operators enter the information incorrectly by transposing numbers. The 546 part of the number is in the same row on the keypad as the 798 portion of the number, so transpositions are difficult to avoid.
This error results in unreadable reports that the inventory accountant must manually correct by going to the shop floor and tracking each job by hand. Obviously, data entry inaccuracy is a big problem in this instance. In the real world, it is an enormous issue because employees are asked to enter data into computer systems even if they are not properly trained in data entry.
THE FACT: Many researchers recently observed a situation where a workforce whose primary language was not English, and which also experienced an annual turnover rate of greater than 200%, was asked to enter production data into a warehouse database; the results were continuing inventory record inaccuracy levels of 50% or greater despite weekly cycle counts.
In short, the human element of data entry can cause considerable difficulty in ensuring that accurate data is entered into a computer database. This problem can be resolved through the use of bar codes.
FYI: A bar code is a set of alternating parallel bars and spaces of different widths that signify letters, numbers, and other characters. When scanned by a laser beam attached to a computer chip containing a decoding algorithm, this cluster of bars and spaces is converted to an alphanumeric character. Several algorithms result in different types of bar codes. One of the most popular is Code 39, which contains both letters and numbers (i.e., is alphanumeric) and is heavily used in manufacturing.
Another is Interleaved 2 of 5, which contains only numeric characters; this bar code is most commonly found in the automotive, warehousing, and baggage handling industries. Yet another variation is the universal product code (UPC), which is primarily found in supermarkets and in the retailing industry. Whatever the method used, all of these bar codes can be generated within a company by entering the required characters into a computer, which converts them to the needed bar code format and sends them to a printer. A laser printer is recommended because it yields a higher-resolution bar code, although inkjet printers are close in comparative levels of resolution. Dot-matrix printers are not recommended for bar code printing because of their much lower resolution levels.
Whatever the type of bar code used, the subsequent processing steps are the same. A bar code is manufactured at the point of use, typically by a special application printer that only produces bar codes. The bar code is typically a self-adhesive one that is affixed to the item to be tracked; this procedure can be automated if the volume of activity warrants investment in such machinery. Then the item being tracked moves through whatever process is occurring and is scanned at fixed points in the process. This scanning can be conducted by a person with a handheld scanner or by an automated scanning station. The scanner extracts information for the bar code and feeds it directly into the computer database.
There are several types of scanners, and the choice of model depends on the application. The main categories of scanners are as follows:
- Light pen. This is the least expensive type, requiring a user to manually drag the scanning device across the bar code. It has a low success rate and may require several scans before an accurate scan is completed. It is most commonly used for low-volume applications where the speed of scanning is not important and where low cost is the determining factor of use.
- Handheld scanner. This device contains a motor that rapidly sends a series of laser scans across a bar code, resulting in a much faster scan. It can also be used with bar codes printed with relatively poor resolution. This scanner can be used with a direct wire linkage to a computer or through radio transmission to a local radio receiver, thereby allowing roving use of the device. A handheld scanner is several times more expensive than a light pen, and radio-frequency scanners usually cost several thousand dollars each.
- Stationary fixed-beam scanner. This device is not intended for manual use. Instead, it is fixed in place at a point past which items are moved, such as on a conveyor belt. The scanner must achieve success on a single scan of any passing bar code or no read will result. To handle this situation, the conveyor belt must be equipped with a shunting gate so that the unscanned items are pushed to one side, allowing machine operators to move them back through the scanning station for a second attempt.
- Stationary moving-beam scanner. This device is the same as a stationary fixed beam scanner except that it is equipped with a motor that sends a series of scans over each bar code, ensuring a high percentage of successful scans. This type of scanner is more expensive than the fixed-beam variety, but its added cost can be offset against the reduced (or eliminated) need for a shunting gate and the manual labor associated with it.
Bar coding is tailor-made for inventory transactions. For example: an inventory identification number is often randomly assigned to a component or product and so has no meaning to the person entering it into the computer system for a transaction. This situation leads to inaccurate data entry. To avoid this problem, bar codes can be attached to all inventory items, which are then scanned as part of any inventory move transaction.
Another inventory-related use of bar codes is shop floor control. As a job works its way through the production area, some companies require the production staff to extract information from a routing sheet attached to the job and enter it into a local data entry terminal. This information tells the production control staff where the job is located in the production process and can also be used by the accounting staff to determine the costs that each job has compiled thus far. It is possible for the data entry person to enter this identification incorrectly, so bar codes can be added to the routing sheet in place of written identification information. The data entry person then scans the bar codes into the local data entry terminal instead of making a typed entry.
Clearly, there are many uses for bar coding. It is ideal for situations where the risk of data entry error is high and is also useful when a company wants to use automation to avoid manual data entry. However, there is a cost associated with the purchase and implementation of bar code printing and scanning equipment, so the inventory accountant should first calculate the costs and benefits associated with the use of this equipment before proceeding to an actual installation.
Wireless Data Transmission
When a transaction is entered into a computer terminal, it travels through a wire or fiber-optic cable to a database for storage. Unfortunately, this data entry method requires one to walk to a fixed terminal location in order to enter data, which is not always possible for employees who collect data as they travel through a facility.
The answer to this problem is to obtain a terminal that sends wireless transmissions to a receiver that in turn is directly linked to a database. This allows data entry to take place virtually anywhere. This mode of data entry has improved rapidly, and several types of portable terminals have been developed. One is the radio-frequency bar code scanner, which is an integrated liquid crystal display, keyboard, and scanner.
It is frequently used in warehouses, where cycle counters can enter quantity changes on the spot rather than write them down, walk to a terminal, enter the data, and then walk back to the counting area. Another terminal is the wireless Palm computer (and several knockoff versions thereof), which one can enter information into with a stylus and then send it to a Web site, from which it is sent as an electronic message to a company’s database. Yet another variation is a portable computer linked to a cellular phone; a modem connection is made through the phone, which transmits data over a phone line to the company, where it is converted to a digital signal and sent to the corporate database.
Wireless applications are directly applicable to inventory transactions. For example, a major problem with any inventory system is that the warehouse staff conducts a transaction and then must find a computer terminal in which to enter the information. This may involve a long walk, so there is some risk that the worker will forget some of the information to be entered or entirely miss making the entry.
Radio-frequency bar code scanners avoid this problem because they are readily available for use no matter where the worker travels within a facility. The information is scanned or punched into the portable unit, and the transaction is immediately sent to the central computer database for updating.
Also, any manager who wants to ensure a high level of inventory accuracy must send an employee into the warehouse to confirm that the inventory quantities listed in the computer are the same as those on the shelves. The trouble is that the cycle counter must plod through the warehouse with a thick sheaf of inventory reports, locate the item to be counted in the report, find it on the shelf, write down any corrections, go back to the terminal, and enter any changes. Clearly, this is a time consuming process. A better approach is to use a radio-frequency bar code scanner to scan the part number of the item on the shelf, scan the bar code for the item’s warehouse location, have the scanner immediately reveal whether there is a counting discrepancy by accessing the central database, and then making a correction on the spot.
Assuming a high level of staff training, the adoption of a wireless system combined with bar-coded transactions can push a company’s inventory transaction error rate to well under 1%. Also, given the reduced amount of time required to enter transactions, one can count on the labor capacity of the warehouse staff to increase substantially.
Radio Frequency Identification (RFID)
A major problem with any manually operated inventory system is the vast number of transactions required to track receipts into the warehouse, moves between bins, issuances to the shop floor, returns from the floor, scrap, and so on. Every time someone creates a transaction, there is a chance of incorrect data being entered, resulting in a cumulative variance that can be quite large by the time a stock item has wended its way through all possible transactions. Incorrect inventory information leads to a host of other problems, such as stock-outs, incorrect purchasing quantities, and a seriously inaccurate cost of goods sold. Although bar coding applications can resolve these problems, bar codes can be destroyed in some environments, are too difficult to read, or require too much staff time to locate for scanning purposes.
One way to avoid these transactional errors is to use the new RFID technology. Although only recently formulated,2 the technology has already been adopted by Wal-Mart, which should ensure a rapid rollout in at least the retail industry. The basic RFID concept has been around for years—attach a tiny transmitter to each product, which then sends a unique encoded product identification number to a reader device. The cost of these transmitter tags has dropped to about 10 cents (depending on their level of complexity and power source), which begins to make it a cost-effective alternative for some applications. Growing use of the technology will likely reduce the cost further.
When a tagged inventory item is passed near a reader device, the reader emits a signal, which powers up the tag, allowing it to emit its unique product identification number. In order to read a large number of tags, the reader turns on each tag in sequence, reads it, and turns off the tag, thereby preventing confusion with repetitive reads. The tag information is then logged into the inventory tracking system, indicating an inventory move past the point where the reader was located.
The most likely implementation scenario for RFID is to first roll it out within the warehouse and manufacturing areas of a company, using it to track entire pallet loads (good for receiving and inventory control transactions) and then implementing it for smaller tracking units, such as cases (good for picking, cycle counts, and shipment transactions) or even individual items (most applicable for work-in-process inventory or retail applications). This implementation approach allows for a progressively increasing investment in the technology as a company gradually learns about its applicability.
A major advantage of RFID is its ability to provide inventory count information without any manual transaction keypunching. This eliminates the need for manual receiving, inventory move, and issuance transactions. It can also provide real-time information about the precise location of all inventory, which can assist with locating missing inventory, arranging cycle counts, and auditing stock. If issued to suppliers, this information tells them precisely how much inventory is currently on hand, so they can more accurately determine when to deliver more stock to the company.
An additional capability of RFID is the activation of an alarm if a tagged item is shifted off the company premises. Another possibility is the use of a more expensive self-powered tag (currently costing about $15 each) that can actively relay its precise location in relation to a fixed overhead positioning unit. The latest tag technology also allows one to rewrite the information stored on a tag many times. Yet another option is to track trailers in a storage area by affixing a single self-powered unit to each one, thereby solving the problem of where specific inventory batches are located. Finally, RFID can be used as an error-prevention device to ensure that goods intended for a specific customer are not loaded onto the wrong truck.
CAUTIONS!: One problem with RFID is the possibility of radio interference, which can be a major problem in heavy manufacturing environments. As a general rule, if wiring in the warehouse and shop area must already be shielded in order to ensure proper data transmission, then RFID may not work. If this potential exists, then be sure to conduct extensive transmission testing in all areas where inventory may be tracked to ensure that radio interference will not be an issue. Another problem is that certain products, such as steel or fluids, obviously cannot be tagged. Yet another issue is that no suppliers have yet developed a complete turnkey RFID solution, so companies are still forced to use consultants to cobble together a disparate set of components into a working system.
Given the difficulty of setting up these systems and the introductory level of much of the technology, it is impossible to install even a simple system for under $100,000, with large multi-site installations costing well into the millions.