How Does Circana calculate CWD % for RMA geographies, where every store and every scan is being shared with Circana? How is it calculated for projected geographies, where sample store data is projected to account for stores which don't share data?
CWD is a distribution percentage calculated by assigning a “weight” to each store, rather than a standard distribution percentage which weighs each store in a market area equally in its calculation. The “weight” in CWD is determined by what percentage of a geography’s entire revenue is being generated by each store. CWD is category-specific calculation, so Circana assigns a weight to each store based upon their total dollar sales for a specific category (the Beer category, the Wine category, or the Spirits category).
Using the Beer category as an example (which includes beer, cider, hard seltzer, FMBs, etc.) Circana calculates the Beer category sales for each store location in a geography for each week. This determines what percentage of the entire geography’s sum total of beer category revenue was generated by each store location. Each store is given a “weight” corresponding to its percentage of the geography’s total revenue. So, if a store generates 10% of its geography’s revenue in a given week, and a 12pk of beer was sold in that store during that week, that 12pk gets 10 percentage points of CWD. If it also sells in other locations being tracked by Circana that week, more points will be added to its weekly CWD %. In geographies which aren’t RMAs (ones which use projection models to account for sales in stores which don’t share data with Circana), a multiplier is assigned to each of the stores which Circana is collecting data from, to project dollar sales (and CWD %) for the entire geography.
Calculating CWD % for RMA geographies (the only geography type which doesn’t utilize a projection model to account for other stores in the area which don’t share data with Circana):
Using the Beer category as an example, Circana calculates the total Beer category sales for each store location in the geography.
Circana’s method for calculating the CWD % for a single item, a brand, or an entire brand family in a RMA geography is fairly simple. RMA’s are census, meaning that every store/every scan is being recorded. Each store represents itself and Circana is not using projection to account for other stores in the area which do not share data with Circana.
So, with RMAs, if one store in the RMA has category sales which make up 5% of the geography's total category sales in a given week, that store has a “weight” of 5. Any UPC which scans at that store during a given timeframe will be credited with 5 percentage points of CWD for that week.
Since CWD is a "Max" calculation, if we're looking at the CWD % for an item/SKU in a timeframe comprised of multiple weeks (e.g., 26 weeks) the CWD we get is the highest weekly CWD. In other words, whichever one of those 26 weeks had the highest CWD % determines the CWD % for the entire time period.
Calculating CWD % for CRMA geographies and all other “sample-based” geographies (all of which utilize a projection model to account for other stores in the area which don’t share data with Circana):
Using the Beer Category as an example, Circana first calculates Beer Category sales for every store in their sample and applies a projection factor to each of them (to project for comparable stores in the geography’s region which are not in their sample).
A projection factor (a multiplier) is applied to each sample store. This is the exact same multiplier that Circana uses to estimate dollar sales in a sample-based geography.
So, if a sampled store is given a projection factor of 5, their beer category sales will be multiplied by 5 to account for the beer category sales of stores in the geography's footprint which aren’t being sampled.
If sample store “A” has recorded $5,000 in beer category sales in a single week, and a projection factor of 5 is applied, that store has a weight of $25,000. Then, by adding $25,000 together with the weekly totals from every other sample store (each multiplied by their own unique projection factors), Circana calculates their projected category dollar sales for the entire geography.
For the sake of this example, let’s say that the projected category sales for the entire geography in a given week adds up to $100,000. Store “A” accounts for $25,000 (with Circana’s multiplier applied), and $25,000 is 25% of $100,000. Therefore, during that week, any UPC which scans at store “A” will be credited with 25 percentage points of CWD in that geography.