facebookexperimental / facebookexperimental/Robyn
Robyn Allocator Same(Similar) budget spend for all models
- Dominant language
- Jupyter Notebook
- Stars
- 1.5k
- Forks
- 433
- PR merge metrics
- No merged PRs in 30d
Description
I've been working with the test dataset provided by the Robyn module and observed an issue regarding the budget allocation across different models. Specifically, I've noticed that the budget spend is very similar, almost identical, across most models generated by the allocator. This uniformity is quite unexpected and may point to an underlying issue in the allocation logic or model differentiation process.
Steps to Reproduce:
1. Load the test dataset provided with the Robyn module.
2. Run the budget allocation using the Robyn allocator.
3. Compare the spend across the generated models.
Expected Behavior:
The budget spend should vary between models to reflect different strategies and optimization outcomes.
Observed Behavior:
The spend across the models is almost the same, which reduces the effectiveness of running multiple models to identify the best allocation strategy.
Note :
I tried running the budget allocation over all solutions as well as clustered models, and the issue persists in both cases.
Thank you for your attention to this matter.
[optimized_spend_results_Clustured.xlsx](https://github.com/user-attachments/files/16560773/optimized_spend_results_Clustured.xlsx)
Contributor guide
Assessment
This issue has not been assessed yet.