• What’s with data identifiers?
• Why is system integration so hard?
• How do I overcome human error?
These are some of the oldest problems associated with new bar code applications. This report shows how three companies solved these common but perpetual problems.
Identifying Data Entry Problems
Draper Inc. maintains an inventory of more than 2,600 different SKUs of window shade fabric for an ever-increasing variety of designs. Each window the company dresses up calls for custom-made shades.
Keeping track of the materials that make shades isn’t easy. Material comes in 30-to 35-yard rolls and each type of material may come in a number of widths. Rolls are stored in 35 different racks, 12 feet high, eight to 10 feet deep. In storage, only the ends of the rolls show.
Pick tickets are printed and workers locate material in the racks, move it to a work cell then, if there is still material left on the roll, return it to inventory. This is an invitation to inventory entropy, a.k.a., disorganization creep.
Data was recorded during the day and end-of-day batch updates were sent to the company’s MAC-PAC XE ERP. But fabric often couldn’t be found when needed; worker’s, supervisor’s and sometimes even sales manager’s time was wasted hunting for material; and promises were being broken because of stock-outs.
Draper worked with Integrated Barcoding Systems, Inc. (IBS) on a data collection solution.
Window shade material is received on skids that hold between 12 and 16 rolls. Protective metal caps cover the ends of the rolls. Inside these caps are bar code labels with Draper’s SKU number, purchase order number and line number of the purchase order. The labels are removed from the end caps and placed on the ends of the rolls.
As rolls are stored, an Intermec CK31 hand-held reader is used to scan the bar codes on the label and a bin location in one of the racks. This information is communicated wirelessly to IBS’s QuikTrac software to instantly update the inventory and location.
When pick orders are issued, fabric is moved to a work center and cut to size. An Intermec SR30 is used to scan a bar code on the work order to record the job number as well as the SKU bar code on the roll. The amount of fabric used is manually entered into the system. Before any remaining fabric is returned to the rack, a new end label with the SKU number and quantity is printed on a Datamax I-series printer. When material is returned to the rack, an Intermec CK31 is used to scan the roll and storage location bar codes.
At the beginning, problems started creeping in as small modifications were made to the initial system when glitches were found and improvements were identified. For example, Draper now uses ISO/IEC 15418 data identifiers (DIs) in their bar codes. DIs are short alpha or alphanumeric strings that precede data in a bar code, indicating what the data represents—SKU number, quantity, location, work order number, etc. This ensures that even if symbols are scanned out of the expected order, the correct data is entered into the correct field.
The more fields on each data entry screen, the more opportunity for errors. That’s why, according to Draper production planner Brian Hickok, operating without DIs can lead to inventory mismatches. The person doing data entry would have to be responsible for making sure data were entered in the proper sequence. That invites errors.
Typically, Draper handles 4,000 to 6,000 transactions (material moves) per month. The company performs cycle counts of approximately two-thirds of their inventory per month. Accuracy is up from less than 80% to 95%. The 5% error is attributed to human error.
Inventory levels were decreased by 25% even after adding a new product line because of tighter inventory controls.
The number of orders that couldn’t be completed as promised because of stock-outs dropped from 30 to 40 per month to fewer than five.
According to Hickok, the system was “easy to sell” in the shop because it saved people time in looking for materials. The time it took to do a physical inventory went from three days with 8 to 10 people down to two hours with three people. And accuracy increased to 99.6%.
His next project will be putting in a bar coding system for miscellaneous supplies such as gloves, racks and drill bits.
“We’ll have to write some tables so when we scan a UPC code it converts to a part number we’ll recognize in our ERP system,” he says. “The challenge there will be the number of items and the table that will have to be written and maintained. A 12-digit UPC code will convert to a 15-digit part number that we have to understand in the ERP system. That will take some doing.”
Why Is System Integration So Hard?
According to Andy Jacobs, president of IBS, a big lesson from many first-time bar code integrations is to involve both the information technology (IT) people and the people who will be working with the equipment.
“Our QuikTrac system rides on top of I-series computer ERP packages,” he says. “Sometimes the project leaders we work with don’t involve the IT staff. Then when they need IT support they say they don’t know anything about the project. Such companies could have gotten their payback a lot faster if they had involved IT on the ground floor.”
In the same way, if IT rolls out a system without involving the system’s users, those users may say “That’s not how we do that transaction out here.”
Jacobs recommends answering the following questions when implementing a data collection system:
• How do you receive today?
• What are the pitfalls and pluses?
• What if…?
“Answering what-if questions may take more time and money up front, but those are the projects we’ve installed where there were far fewer headaches and problems because all those issues have already been addressed,” Jacobs says.
A Project by the Books
Powell’s Books, a bookstore chain based in Portland, Ore., has two warehouse locations storing an average of 400,000 books. Warehouse workers must keep track of nearly 10,000 shelves.
Workers would scan a book’s SKU number and have to enter the locator code during the data entry process. While data entering they would keep a tape measure with them to measure their stack of books to see if they exceeded the amount of space on the shelf. When they hit their limit they would then have to manually change the locator code in the data entry program and keep tabs on what books belonged on what shelf. This putaway method was extremely time consuming and inefficient.
Powell’s Books deployed 50 Psion NEO rugged mobile devices in January 2010. The new system integrates both bar code scanning and SKU scanning, making it easier and more convenient for workers to create a location for each book. The NEO is also able to perform cycle counts each year to track inventory.
Powell’s Books has been able to save two to three hours on data entry. Although more staff members are using the devices to stack books, time savings on data entry has been a huge benefit. Additionally, the NEO devices have enabled the cycle counting process to be 20 to 30 percent faster, more accurate and staff members believe they have spent less time going back and forth to the shelves to fix or add books.
Jason Ellingson, used book distribution manager for Powell’s Books, does point to some important lessons from this project, however.
“Not all the scanners would scan the same code the same way,” he explains. “That’s more a matter of software programming. During testing we’d get these new bar codes and if one person had one type of scanner at their computer it might scan fine, but at a different computer with a different scanner it wouldn’t scan as well. In the warehouse we scan books out when we transfer them. Some of those computers have different scanners on them. You need to make sure all the equipment performs the same way.”
Account for Human Error
A switch supplier for the automotive industry also makes window lifts, power mirrors and steering wheel switches, among other electronics. The company’s warehouse stocks thousands of components, many of them small and often similar in shape but different in size and function.
When a part was needed, pickers would go to the computer, enter the part number, and be directed to the pick location. Once there, it wasn’t unusual for the picker to look down to scan the product bar code, enter the number of parts, then look up and reach into the wrong bin.
This supplier worked with William Frick & Company, supplier of labels, tags and signs, to develop bar code labels in colors that corresponded to newly color coded warehouse shelving. These color coded shelf locations were programmed into the company’s inventory database along with product numbers and aisles. Each bin was given a colored label with product name, number and bar code. Now pickers can easily distinguish needed products by the color coded shelf on which it is stored.
As a result, this switch supplier improved inventory control and accuracy. They had more efficient line replenishment with fewer mistakes by warehouse personnel.
Reducing the errors associated with bar code labeling, programming and scanning not only improves efficiency, but it raises morale—which will also reduce mistakes and increase worker interest in fixing any others that crop up later.