
Magic™ Computer Generated Ordering
Optimized Inventory:
Every store. Every item. Every day.

Itasca Retail pioneered store-level, real-time, computer-generated ordering software with its first deployment in 2003 at Price Chopper Supermarkets. It delivered superb returns then and it continues to lead the industry, delivering better than 99% on-shelf availability for our retail customers. Magic™ has been successfully deployed in more than 2,900 supermarkets across North America with more being added every day.
Optimize...
Magic™ CGO is a proven software solution that optimizes replenishment and management of store-level inventory and returns astronomical ROI. This ensures the right amount…at the right time.
...in real time.
Inventory levels change…in the blink-of-an-eye. You often don’t know how fast until you look at the shelf and an important item is missing. We don’t let that happen to you…ever.
Why Magic?
Our valued customers enjoy…
- the highest increases in sales,
- the largest decreases in inventory,
- and the best inventory performance
in the grocery industry.
Here’s Price Chopper to tell you more…
Blind Orders
Aren’t “blind” anymore. Your inventory system knows the amount of product en-route…and considers it when producing an order.
Perpetual Inventory
You’ll always know exactly what’s on-hand (or on-the-road for that matter) in every department in the store.
Pre-Book Ordering
You're likely thinking (because almost everyone asks)...
Can Magic data/orders be used with manufacturers?
Yes. Magic™ produces "Forecasted Orders" which are actual orders (see above on Forecast vs. Order) with the difference being they are in the future. So, a Forecasted Order uses current information combined with predicted information to derive the most accurate future order possible. These "orders" are automatically stored in our Central Data Server location, usually in a HQ-based server, and can be used if store connectivity goes down. They can also be sent directly to suppliers as advanced notice of store-level demand. We have a customer using them for the packing of store-level orders for fresh product directly at the product origin - this saved 5 shelf days in the products' logistics.
How accurate is your forecast? To what statistical level?
This is an interesting question, and once all the particulars are internalized well, this question becomes almost moot for several reasons. Forecasts are, by definition, always inaccurate, because even the best, most accurate forecasts aren't 100% correct, 100% of the time. So the question becomes: How can we minimize the error rate considering all factors? We are afforded a bit of a reprieve here in that our forecast (and really the order itself) only have to be as accurate as a case-pack: the smaller the case-pack, the closer to the actual unit forecast the better. But in the case of the 12 unit case-pack, if the order is "off" by one or two, it's still within the quantity of the case-pack, and we likely would have ordered another case anyway.
What’s the difference between a “forecast” and an “order?”
A forecast is simply the anticipated number of units that will be sold, during a very specific period of time. It is input to the algorithm and actual order that is sent to the vendor.
An order is the amount of product, rounded to the case pack for the product in question, that will be requested from the appropriate supplier. It considers such important factors as Consumer Demand (the Forecast described above), Current On-Hand, Current On-Order, Anticipated OOS (if applicable), Propensity of creating an OOS by NOT ordering compared to the propensity of creating back-stock by ordering.
How do you consider seasonality in your predictions?
Yes. Seasonality must be considered, but it might not be how you think it would be. Too much leveraging of the previous year's data to predict this years can be fraught with error.
Said a different way, "Last year's data is so, well, 'last year.'"
This is because so many things differ from one year to another: some Holiday dates change; multiple products exit, enter and change packaging; prices change; trends begin and end; etc. So, Itasca employs a fast-reacting algorithm to understand the true demand impact of new and changed items as quickly as possible which reflects the most current activity and enables the CURRENT seasonality to be appropriately considered. In addition, we derive from history and ask merchants to verify a list of items that have unique behavior in relation to a particular Holiday. Great examples of these types of items are whole fresh turkeys, cranberry sauce, pumpkin pie filling and stuffing at Thanksgiving time. These items have uncommon sales levels (compared to regular day-to-day sales) leading up to the date, and then drop to almost nothing from then on. The challenge then for an inventory system is to have enough stock to meet demand in the lead-up, but then be back to normal levels so that multiple cases or pallets aren't left in the stock-room, potentially there until the next Holiday. With this list and the historical information, we can accurately predict demand and stock-levels pre-Holiday, and know when to slow-down and then stop further ordering to meet the "normal" day-to-day needs post-Holiday.