Setting Priors
Overview
Rockerbox uses a Bayesian methodology in our marketing mix models. An important step in building models is setting priors. Setting priors in Bayesian models allows for the incorporation of prior knowledge or beliefs about parameters before modeling. Carefully chosen priors enhance model robustness and improves estimation accuracy.
For MMM, Rockerbox sets priors for a given channel or subchannel based on information we acquire from different sources, including:
- In-platform performance metrics, such as Facebook's reported ROAS
- ROIs from alternative measurement tools, including MTA
- Incrementality or lift studies
Customer-Provided Performance Metrics
If you have channel or subchannel-level ROIs—from external reports, incrementality studies, etc—provide them to your Rockerbox CSM to include in modeling. If you don't have any other performance metrics, or for any channel where you do not provide these metrics, then Rockerbox will use defaults based on the closest match to the channel or subchannel.
Performance Metrics Templates
Use this template to quickly document the set of ROIs for each of your channel or subchannels. Make a copy, replace the sample values with your own, and deliver it to your CSM.
Example
Channel | Subchannel | ROI |
Prospecting | 2.5 | |
Retargeting | 4.2 | |
The Trade Desk | 1.8 | |
3.1 | ||
AdWords | NonBrand | 2.2 |
AdWords | Performance Max | 3.5 |