By Jim Lewis, CEO Enhanced Retail Solutions LLC

I’ve been in the retail and wholesale inventory prediction business for a long time. A lot has changed. While our tools are light years better than when I started (big shoulder pads were a thing) what hasn’t changed is that some items still require fine tuning regardless of how much you paid for a forecasting system. Think of it as the 80/20 rule. But what if we can improve that to 90/10?  Or 95/5 (I’d say 100/0, but I am a realist). When inventory productivity improves it has a direct relationship with sales and profit. That’s why we’re continually enhancing our tools to meet those goals.

Getting to Know An Item

Selling behavior can be very diverse even for items within the same product category. That’s why we developed a system that really gets to know each item. It sizes it up so to speak. How mature is it? How consistent is its selling? How promotional is it? Is it seasonal or year-round, etc.? If the forecasting tool can assign a more appropriate and accurate model to forecast the item, that’s less time someone needs to fine tune it. It takes an inventory planner’s capability to a whole new level.

Classifying items based on age, seasonality, consistency in selling and other statistics help determine which planning model to use.

Determine the Most Appropriate Approach

Once the system understands how the item behaves, and assesses confidence in its history, the most appropriate model is selected. It’s smart enough to understand if you are out of stock it needs to add lost sales back into the basis. It adjusts the seasonal profiles every week based on the latest trends.

Sometimes it will use a very short historical period to get rate of sale. Other times it will take a longer period. Or for extremely low volume items, it may just match sales from the previous year or apply a company trend to it.

There’s not just 1 thing going on. It is essentially creating the logic needed for each item based on the multiple classifications and rate of sale. An item can have intermediate volume but be somewhat seasonal. There are multiple levels within each classification. For example, seasonal quarterly or half yearly. It can be sporadic or spread but mostly year-round, etc.

The system automatically assigns and shows which model it chose to plan each item.

Seasonal Profiles

A seasonal profile creator runs every week and is curated for every item. It’s a smart system that looks at the consistency of sales for the item and then determines whether to create the profile based on its own history, or if a higher level, such as subcategory, category, brand, etc. makes more sense.

Once the system makes all these determinations, it estimates future sales and then lines up the inventory and on order against it. Then based on other parameters such as lead time and weeks of supply, recommends needs and when to place them with the factory.

This approach to inventory planning is very practical and much smarter than many of the overpriced familiar branded systems. It’s never been brain science, and to me, getting that “feel” for an item integrated into the forecast is a win-win.