Open eakl opened 7 years ago
Original Post by nicogla
I see two ways to estimate synergies effects:
1) Use a regression with interaction effects between copies. If copy1 is a dummy variable (0/1) for the households exposed to copy 1, and copy2 is the same but for copy2: purchase~copy1+copy2+copy1*copy2 will estimate the effects of copy1, copy2 and the synergy when you are exposed to both copy1 and copy 2 respectively.
2) An alternative method based on AVI score? I would do the same as an AVI but for the households that have been exposed to both copy1 and copy2. Then I would compare that AVI with the “single effect” ones. If AVI_copy1&2 is significantly larger than the marginal effect of both AVI_copy1 and AVI_copy2, you have your answer. E.g. if AVI_1=120 and AVI_2=135, then AVI_1&2 needs to be larger than 155... But I’d be curious to know how Mars does it exactly…
Original Post by max_metral
How can we calculate a possible synergy between copies? To see if some combinations are better than others, i.e. 1+1=1 or 1+1=3.