# Suitable and efficient comparison metric/index for price data

I have price data for similar products in different stores like below

productId   StoreA  StoreB  StoreC  StoreD
00001       14      10      14      15
00002       35      34      36      40
00003       100     94      90      94
00004       14      14      25      14
00005       11      15      13      10


StoreA is my base store to which I want to compare other store's prices. Currently my approaches are naive.

Approach 1: Calculating %change = (storeA/AVERAGE(storeB, storeC, storeD))-1 # MEAN is sensitive to outliers so in ID 00004 all prices are same in all stores except storeC so overall I am not getting correct metric/index which represents the data

Approach 2: Calculating variance = ((storeA - storeB)**2+(storeA - storeC)**2+(storeA - storeD)**2)/3 # Here storeA is my base store so I am taking it as mean price


Any better approaches to get a single index for comparing prices with a base price ?

• Two methods make sense to me, but it actually depends on what you are trying to achieve. What is the index going to to be used for? Can you elaborate on what problem are you trying to solve? May 17, 2017 at 8:55
• The simple objective is to determine that storeA on average is performing better then all other stores or not. May 17, 2017 at 9:01
• If better means better than all other stores or not - then per each product i would check if indeed StoreA meets this term: I.E - StoreA >= StoreB and ... May 17, 2017 at 9:35
• Do you also have sales volumes? If very different maybe you need weighting. Sep 30, 2019 at 13:15