Evan Schuman's StorefrontBacktalk

One of the worst parts about managing retail businesses is dealing with unknown future returns. Is that booked revenue all real? You can certainly know that, statistically, XX percent will be lost to returns. But is it possible to know more specifically?

What if your system could look for hints about specific purchases that could be flagged for likely returns? Perhaps a customer who purchases three of the identical shoe, but each one in a slightly different size?

Or perhaps any purchases from a customer who has a history of high returns? What about some social site commentary that specifically flags an imminent return? This topic cropped up in a conversation this week, and we wanted to throw it out for somebody’s functionality wish list.

One key attribute would be the ability for a system to accurately project likely returns. We’re not talking about a generic prediction, such as 5 percent of this SKU will probably be returned. Rather, we’re looking to project specific customers returning specific merchandise within specific date ranges. How could this be done? Quite easily, actually.

Beyond watching for the multiple-unit-similar-sizes purchases, why not examine CRM profiles and purchase/return history? Your top serial returners generally have established patterns, and you can accurately project how long they will typically wait before processing the returns.

Could this system ever be accurate enough to impact inventory needs? If there’s a product you think you will just barely have insufficient stock of, wouldn’t you like to be able to query the system to see if any of that SKU are likely to be returned shortly?