Ensuring accurate inventory records every time
Your information system and the plans, recommendations, and alerts it gives you are only as good as your data.
The most fundamental data reveals inventory balances and availability. It is generally accepted that inventory records should be at least 95-to-98 percent accurate in order to enable successful enterprise resources planning and provide good customer service. While many companies large and small are able to achieve and maintain this level of accuracy, many more continue to struggle with this requirement.
The APICS body of knowledge emphasizes that cycle counting is the key to accuracy. By counting items regularly and using the cycle count process to identify and correct the sources of errors, companies can quickly raise accuracy levels beyond 90 percent and maintain a continuous improvement posture that maintains a high level of performance.
However, cycle counting alone does not improve inventory accuracy. Follow-through to identify and correct the source of errors is essential before improvement can be attained. For example, transaction errors, or inventory movement that is not reported quickly and accurately to the system, result in incorrect inventory records. Merely correcting the balance essentially treats the symptom but ignores the disease. Obtaining a high level of accuracy requires preventing the errors from occurring in the first place.
Machine learning
What’s the secret for achieving accurate inventory records? Well, it’s not exactly a secret: The only way to achieve and maintain a high level of inventory accuracy is through a solid and reliable transaction reporting process. It’s not fun, and it’s not glamorous—but it is essential.
A transaction can be created by a person or a machine. Machines, of course, are a lot more reliable. Thus, any transaction reporting that can be automated should be automated. If a transaction report is created by a human but can be computer-assisted or computer-validated in real time, it is certain to be more accurate and reliable.
The following represents the hierarchy of transaction reporting, from most to least reliable. Use this information to maximize your inventory accuracy. The most accurate and reliable are fully automated systems. These are frequently directly connected to sensors or programmable logic controllers to ensure proper reporting of inventory movement and usage. The next level of accuracy is computer-assisted or computer-validated entry. This includes the following technologies:
Radio frequency identification: This is similar to direct connection, but errors can be introduced by signal interference or bad scans.
Bar codes: This technology also is subject to bad scans and connectivity issues. There also is the possibility of people scanning the wrong codes or incorrectly entering non-scanned information, such as counts. Computer systems can validate much of the data, including item, location, and order number.
Validated manual entries: Most systems will authenticate at least some of the data that is entered by humans. Computers can be configured to detect whether an issued item is actually on the bill of material or pick list, if the quantity is as expected, if the item really is in the scanned location, or if the lot number is correct, among other functions.
Finally, the least accurate method is manual entry. Unfortunately, systems often cannot validate miscellaneous or unexpected transactions such as adjustments, unplanned issues, scrap reporting, and unauthorized uses including samples for testing or marketing giveaways. Further, these transactions are unlikely to have bar codes except possibly for item number and location.
Only a disciplined and reliable transaction reporting process can help transaction errors from occurring. Cycle counting helps, but it’s not enough to achieve desired levels of accuracy. To achieve your goals, it is necessary to make the best use of automation and error correction.
Enterprise Insights by Dave Turbide, reprinted with permission from APICS magazine | July/August 2014 | 24 | 4