Robyn is an experimental, AI/ML-powered and open sourced Marketing Mix Modeling (MMM) package from Meta Marketing Science. Our mission is to democratise modeling knowledge, inspire the industry through innovation, reduce human bias in the modeling process & build a strong open source marketing science community.
I’m encountering some unexpected behavior with the Robyn Budget Allocator and need your assistance to understand what might be going wrong. Here are the details of my issue:
With Max_Response:
Spend: $1M
Response: $5.42M
ROAS: 5.42
With target_efficiency (target variable ROAS = 3):
Spend: $2.38M
Response: $7.14M
ROAS: 3
For the same time period, my initial values before allocation were:
Spend: $855K
Response: $3.34M
ROAS: 3.91
Additionally, when using Max_Response, the budget allocator suggests a few channels, which seem to be based on mROAS. However, when I switch to target_efficiency, it significantly decreases the budget for these initially suggested channels and instead recommends a different channel, allocating 75% of the budget to it. The channels suggested under Max_Response align well with the business team's insights, making this shift even more puzzling.
Issues observed:
The spend and response numbers increase significantly when targeting a lower ROAS (3) compared to the initial ROAS (3.91).
When setting the target ROAS to a value greater than 3.91 (e.g., 4 or higher), Robyn suggests decreasing the spend, which makes sense. However, the behavior is inconsistent when the initial_ROAS > target_ROAS.
There is a dramatic shift in budget allocation between Max_Response and target_efficiency, suggesting different channels and budget distributions.
This behavior is confusing and seems counterintuitive based on my initial values and business insights. Any guidance or explanation for this issue would be greatly appreciated.
I’m encountering some unexpected behavior with the Robyn Budget Allocator and need your assistance to understand what might be going wrong. Here are the details of my issue:
With Max_Response:
With target_efficiency (target variable ROAS = 3):
For the same time period, my initial values before allocation were:
Additionally, when using Max_Response, the budget allocator suggests a few channels, which seem to be based on mROAS. However, when I switch to target_efficiency, it significantly decreases the budget for these initially suggested channels and instead recommends a different channel, allocating 75% of the budget to it. The channels suggested under Max_Response align well with the business team's insights, making this shift even more puzzling.
Issues observed:
This behavior is confusing and seems counterintuitive based on my initial values and business insights. Any guidance or explanation for this issue would be greatly appreciated.
Thanks!