Spaces:
Runtime error
Request for Additional feature
Can you show how we can adjust dynamic pricing with multi product and its regressor graphs how acutal price and adjusted price generates more profits
Hi,
For the multi-product case, the issue is that the pricing of product A can influence product B because they can be (partially) substitutable (i.e., "cannibalization").
We can differentiate between two situations:
(1) There are few products (e.g., 5-10 products):
Here we can reformulate the univariate demand function D(p) = f(p) where f(.) is some suitable functional form to the multivariant equivalent D(p1, p2, ...) = f(p1) + f(p2) + [...] + g(p1, p2) + [...] where g(.) accounts for cross-product correlations.
Similarly, the profit formulation becomes something along the lines of Profit(p1, p2, ...) = (D(p1) - Cost(p1)) + (D(p2) - Cost(p2)) + [...] instead of the univariate Profit(p) = D(p) - Cost(p).
(2) There are a lot of products (e.g., hundreds, thousands or more products):
The above is not scalable to cases where there are lots of products. Here we can group cluster substitutable products into "baskets" with a basket price that is the average of all products in the basket.
We can then perform dynamic pricing on the basket prices instead of the individual product prices, either as we did above (if there are few baskets & you want to account for correlations between baskets) or for each basket separately (then you assume that products in different baskets are not substitutable). We assume that the product prices always remain a fixed percentage of the basket price.
Hope this helps!