A database fact/measure that you should not sum over time or market or product is often called “non-additive”. This is a common database term, not specific to CPG data. Dollar and unit sales are both additive. By contrast, distribution and price are non-additive. For example, if one upc has 30% distribution and another upc also has 30% distribution, the distribution of the two items combined is not necessarily 60%. For some non-additive measures (like price), you can calculate an summation correctly by summing the numerator (dollars) and dividing by the sum of the denominator (units). But for some non-additive measures (like distribution), you don’t have the information you need to do the aggregation. You would need to know what was happening at the store level to get it right.
Is TPD a non-additive fact?
As long as you are looking within one geography in one period, then you can add TPDs across products. You cannot add TPDs together from different geographies (markets or retailers) or across time periods.
Hello,
I know that ACV is not additive, but is there a way to calculate total ACV for a Brand when you have 5 UPCs with ACV measure? Can you average the ACV of the 5 UPCs for total Brand ACV?
If you only have the items but not the brand %ACV, then the most conservative assumption for the brand %ACV is the value for the max item. If you take the average of the items it will understate the brand ACV since the brand distribution is always at least as high as the best item, and usually higher. If you want to be a little more accurate I recommend adding about 5% (not 5 points) to the max item and calling that the brand distribution. Of course if that results in a number over 100, then the brand distribution is 100. See point #5 in this post.
Hope this helps!
is no of Buying Households an additive measure?
Unfortunately, no it’s not additive. If you had 100 HH who bought flavor A and 75 who bought flavor B of the same brand, there’s no way to know how many HH bought the brand at all (if you only have item level data). It could be as high as 175 (if all HHs only bought one flavor) or as low as 100 (if all HHs who bought flavor B also bought flavor A).
Thanks for the prompt response! Your website is very helpful 🙂