Hello,
I have launched a CBC study for an FMCG company and there's the doubt of calibrating the shares of preferences to the real market share.
After reading several posts in this forum and this article (https://sawtoothsoftware.com/resources/technical-papers/external-effect-adjustments-in-conjoint-analysis), we have come to the conclusion that it is better not to calibrate to market shares if we can't explain big changes in elasticities.
However, if we use shares of preferences (SOP) and translate changes of SOP to actual volumes to estimate impacts, we observe that the market size changes, due to the fact that the SOP and real market share are not the same. Let me put an example:
Imagine there are 3 products and the SOP actual and after price changes are
P1 - 20% --> 30% (delta SOP = 50%)
P2 - 30% --> 40% (delta SOP = 33%)
P3 - 50% --> 30% (delta SOP = -40%)
Now, imagine that the current volumes in units and hence market shares are:
P1 - volume = 1.000 - market share = 43%
P2 - volume = 550 - market share = 24%
P3 - volume = 750 - market share = 33%
Total market size is 2.300. If we apply the delta SOP to the actual volumes, we obtain the following volumes and growth vs base volumes:
P1 - 1.500 (+50%)
P2 - 733 (+33%)
P3 - 450 (-40%)
The total market size has increased to 2.683 units. This is going against the principle of keeping market size constant. One thing that we could do is to rescale everything to keep market size constant to 2.300 units. If we do this by multiplying all volumes by the original market size and dividing them by new market size, we obtain the following absolute volume and growth vs base volume per product:
P1 - 1.286 (+29%)
P2 - 629 (+14%)
P3 - 386 (-49%)
The issue that we see now is that the volume growth by product is not the delta SOP anymore, and this is affecting the resulting elasticities. That is, the original elasticities obtained from delta SOP is not the same as the resulting elasticities after rescaling the product volumes to keep the market constant.
At this point, we have the following question, what is it better or "more fair" to assume:
- apply delta SOP to volumes and hence allow the market size to change or
- apply delta SOP to volumes and rescale the resulting volumes to keep the market constant but changing the final resulting elasticity?
Thank you for the help!
Kind regards,
Javier