It depends on what you want to measure. If you just want to measure the overall number of transactions or revenue generated by the business, one model is enough. If your goal is to measure the performances of each product then you'll need to build multiple models.
How do you account for all of these changes in the model? e.g. let's say if 40% of products ran out of stock within a 2 month period. They were replenished back to stock in three days which caused a huge surge in revenue. This really disrupts the excel model provided in this course. Thanks!
You can have a new feature that describes the availability of your products daily. Use that feature to describe the impact that there is having a 40% availability of your products. Generally there is no rule that it's always right, but is a game of trial and error until you find the model that suits your needs the most.
If there are multiple products, do we build a MMM model for each product?
It depends on what you want to measure. If you just want to measure the overall number of transactions or revenue generated by the business, one model is enough. If your goal is to measure the performances of each product then you'll need to build multiple models.
Good content, man. Do you have a video showing the final result of this type of analysis?
Hi Andrey, thank you for the comment, and appreciations.
We will start sharing our case studies soon with all our process and outputs
How do you account for all of these changes in the model? e.g. let's say if 40% of products ran out of stock within a 2 month period. They were replenished back to stock in three days which caused a huge surge in revenue. This really disrupts the excel model provided in this course.
Thanks!
You can have a new feature that describes the availability of your products daily.
Use that feature to describe the impact that there is having a 40% availability of your products.
Generally there is no rule that it's always right, but is a game of trial and error until you find the model that suits your needs the most.