Continuing from part two (What are UPC codes), the problem with telling you how to create SKUs is that it can be a matter of opinion, “whatever works best for you”, or whatever your customer says it should be. The definition is very squishy and ambiguous. Because we can’t agree on what information should be incorporated into a SKU, retailers may tell you what to do. For example, Amazon directs their vendors to issue SKUs that incorporate the UPC -but what do you do if you don’t have a UPC? This site (interesting by the way) says that a SKU “is a number assigned by the retailer to keep track of the type, color and size of a product”. As far as I’ve known, a SKU is assigned by the manufacturer to which of course, a retailer could append their own unique identifier. This is specific to smaller stores, a larger entity would require a UPC.

How to issue SKUs
You can do what you like or whatever your customers tell you. Barring those two, the guidelines below will make it easier to transition to UPC codes if you need to later on. As a refresher from the first entry, there is a hierarchy to style numbers, SKUs and UPC codes. UPCs are based on SKUs; SKUs are based on style numbers and how you design style numbers is anybody’s guess (in spite of my best efforts, heh). Here is a little worksheet showing how I’ve designed SKUs. I’ll explain why I did it like this as we go along.

You’ll notice fabric is omitted. You probably don’t need to list fabrication in your SKU because it’s usually covered by colorway but better if by style (number). For example, how often or likely is it that you will sell an identical style number of a given dress in red linen and red cotton? Usually, you sell a style in red linen and blue cotton or something like that. As a practical matter, it is more likely that the patterns of the linen and cotton would be different (based on shrinkage) so they would have a different style number.  Keep this as simple as possible. Perhaps this exercise will illustrate why I think it’s best to assign fabrication according to style (number).

If you’re going to design SKUs, you may as well go full bore and design the system to facilitate future EDI and fulfillment. It is set up a bit differently because UPC and SKU codes may be used depending on the situation. For example, it is likely that only the UPC will be used on the price tag but SKU and UPC will need to be on the carton.

How to issue UPCs
Technically, you can only assign IRN or Item Reference Numbers of the UPC code but if I wrote “how to issue IRNs”, nobody would be very interested or care about it, not understanding how it all ties in together.

Examples: For EDI purposes, your style number must be five digits. If it is shorter, you must “pad” it with leading zeros as in my example above. The color number must be three digits, again padded if it’s not long enough. If either or both the SN/color digits are too long, you will need to redesign your style numbering and/or color coding system (nag). The size must be three digits with trailing spaces rather than padded zeros. There’s more in chapter 11 of The Vendor Compliance Handbook which you should have anyway. While it can’t possibly answer every question, I strongly recommend it (pt.1, pt.2).

If you will be using UPC codes or plan to, the next step is to assign a unique product code to each SKU. As I mentioned before, in UPC parlance, the product code is called the IRN or Item Reference Number. The IRN should be a sequential list and not tied to product attributes because you’re not going to have much choice in the matter. Depending on how many UPC codes you buy, you will be assigned the use of 2 to 5 digits. The fewer the number of codes you buy, the fewer the digits you can use. For example, if you buy 100 codes, you’re limited to three digits. I do not know if the IRN must be padded but I illustrated it that way in the spreadsheet screen shot.

Next: Why you need your own bar codes