Optimizing your marketing channels:
How Marketing Mix Modeling solves attribution and privacy challenges

Choosing the right marketing channels is the eternal problem for all marketers.

Continuously updated budget allocation recommendations

Measure incremental sales from marketing, and marginal returns at different spend levels

Easy sales forecasting and scenario planning

Built-in verification and monitor model performance

No privacy issues

There is no tracking of customers in any way.

Works for all channels

Works just as well for offline channels as for digital marketing, and gives a holistic view of your marketing performance.

Actionable

Get recommended channel mix and simulate forecasted effect when planning your marketing activities.

What is Marketing Mix Modeling?

The goal of Marketing Mix Modeling (MMM) is to help you allocate your marketing budget across different media channels to improve sales, customer acquisition or any other business KPI.

Using historical marketing activities, the model estimates the efficiency of each channel and how it impacts your KPI.

Odins' approach

At Odins we are developing modern Marketing Mix Modeling with a focus on:

Continuous updating and evaluation‍
New data is integrated into the models as soon as it becomes available. Your model is always up-to-date and evolves with changing market dynamics.

Historical model predictions are evaluated against what actually happened, allowing you to build trust and confidence in the models.
‍
Bayesian methods and causality
The Bayesian approach allows for more complex modeling and offers two distinct advantages over traditional modeling:

1. Allows for incorporating prior expert- and domain knowledge into the models, leading to more accurate and stable models.

2. Uncertainty quantification is natively a part of the modeling through probability distributions, giving uncertainty estimates for all predictions. This is vital for the decision making process, and can help inform the design of experiments.
‍‍
By treating the models as causal inference tools, and not just prediction tools, we estimate the true causal relationships between marketing spend and business outcomes.