Lately I hear a lot of people talking very "knowledgeably" about the quality of scanner data. These people lament with absolute certainty, "If we could just disable the multiply key!" I feel compelled to set the record straight. The multiply key is probably 5% of the problem with scan data.
Let's look at the reasons why they multiply: because products don't scan, because one of a group of similar products isn't on the file or has an incorrect price, and because they believe it's more productive to multiply. In reality, it takes cashiers longer to lift multiplied products over or around the scanner than it takes to scan each one.
Multiplication is an issue of symbol quality, file maintenance and training. If the symbols were printed with high quality, item maintenance were handled correctly and cashiers were trained properly, there would never be a need to multiply. And cashiers need to understand that the data will be used, and what it will be used for. Personally, I think symbol quality, file maintenance and training are management's problem, not the cashiers'.
That settled, where do the majority of the problems lie? In three other areas: the in-store scan coordinator or front-end manager, the corporate buying departments and the corporate management information systems departments.
In-store "data damaging" activities include failure to execute batches of price changes and new items on a timely basis. If prices are wrong or items aren't on the file, cashiers will key in prices (or multiply using a different "like" product). Other store-level data damagers: managers change prices during the week and change them back at the end of the week, so you can't tell what caused that apparent movement spike. They also reset movement fields to zero so they can track how much of the product sold at their reduced price, and at the end of the week we only know what sold from the time they reset the field.
Corporate buying departments enter price changes and new items too late, so they don't get into the store maintenance on schedule. They run a sale on multiple flavors of the same product, i.e. cake mix, and they leave one flavor out.
Retailer MIS departments spend too little money on communications lines, and schedule transmissions too tightly. If a store gets missed, or a line is dropped mid-transmission, the data from the store doesn't get captured that week, and the following week we see two weeks worth of data. Corporate MIS departments often don't pull back prices from the stores. When that's the case, prices that are logged in corporate systems to reflect what's in the stores often are incorrectly overlaid onto store-by-store movement.
Those are highlights of some of the areas. There are a lot more. Two Food Marketing Institute documents: "Scan Data Quality" and "Price Maintenance" explain it all in detail, if you're interested. It's easy to blame the cashiers and the infamous multiply key, but trust me, they are the least of our problems.