Applying Avatars to Historical Data

To get value out of creating Avatars, you have to apply them to historical data. Like it or not, an form of planning system relies to some extent on historical data. If you are going to plan at the Avatar level (and it's worth doing) then you need history to reflect that variable. Not as hard as you think. Yesterday I noted that you may have fields in the data record you no longer use...and which can be "reassigned" to Avatar use. This can be done to past data too. It IS possible to run a quick program to go back and edit any form of data repository....at least most of the more modern systems allow this potential.

The first thing to do is to have someone start to assign merchandise to these Avatars based on the one it is most likely to have been bought by....again....if possible...try to validate by spending some time with long-time floor salespeople who can probably remember last year or the year before's assortment. Other validation input comes from the vendor themselves. Who was the item designed to appeal to? Someone at the vendor, assuming they are still there, can probably answer that question. I hope. Do a quick review of any advertising that ran for both the brand and for the specific item. Sometimes a brand is so poorly defined it starts to be a hybrid of two or more Avatars. Bad news when this happens. If possible, try to remember which of your competitors (or non-competitors) also carried that product or that brand. Sometimes your assumptions about what merchandise appeals to which Avatar can be invalidated based on which of your competitors did well with it or which didn't. If someone you really don't want to compete with had the same product/brand and did well with it, you might be wrong about your Avatar....or even about the desirability of merchandising TO that Avatar.

I digress. The point here is to try to make an educated classification of which Avatar an item should have appealed to....or to the Avatar NONE OF THE ABOVE. This is a really important Avatar for several reasons. First, in the new world, you don't really want to be spending scarce open to buy providing merchandise which appeal to NONE OF THE ABOVE. Second, it will allow you to actually critically analyze the NONE OF THE ABOVE merchandise. Be prepared for some shockers. You may actually have done better with some of this than you did for something which seemed right down mainstreet for one of your Avatars. If this happens....GOOD! It means you might have another Avatar you can merchandise to, or a better Avatar. Remember, the more Avatars you have, the less likely you are to make meaningful merchandise statements and assortments. And yes, this really does apply to hardlines. Let me say that again. Yes, this really does apply to hardlines!

Hands on, what you've now done is to restate historical data so that you have sales, markdowns, GM and inventory based on Avatar groupings. If you do this at the SKU level, it can be made to flow across the store level by your existing merchandise system. In other words, you now have a way of seeing what percentage of inventory, sales and gross margin was contributed at the store level by each Avatar. Doing so will allow you see where a store was over-inventoried, under-inventoried or simply ignored from an Avatar perspective. When we did this for one major retailer, the results were extraordinary. Some stores were simply denied the opportunity to do much business with a given Avatar because they were allocated way too much for the others. Adjusting the inventory to reflect this opportunity raised overall sell through for those stores significantly. True.