Marketing Mix Optimization: 4-Month Implementation

Tiger Analytics’ Marketing Mix Optimization solution is a cloud based analytics solution which helps organizations to consolidate, visualize and comprehend historical marketing spends. It helps marketers understand the contribution of each marketing channel towards sales and provides them tools at fingertips to evaluate multiple scenarios to optimize investments across marketing channels.

The solution comes with prebuilt foundational components across Data Engineering, Model building and end user consumption which would seamlessly work on Azure PaaS components providing a comprehensive view on Marketing Investments. Below are the key steps in the workflow of the solution implementation

1. The solution integrates with existing environment and marketing systems as well as external feeds to bring data across media (traditional, search, display, social, CRM etc), non marketing variables & external/control variables using Azure Data Factory and push it into Azure Data Lake Storage Gen2/Blob storage
2. Data is preprocessed using ADF templates and accelerators available as part of Tiger Marketing Mix Optimization solution and stored in Azure Data Lake Storage Gen2.
3. ML models are built and trained in custom Python + Azure Machine Learning using custom algorithm available as part of Tiger Marketing Mix Optimization solution
4. Azure VMs/Azure Databricks is used for computation
5. The ML Pipelines leverage Azure Functions for serverless compute and Azure Load Balancer for distributing traffic.
6. Access Control and authentication for data and the ML workspace leverages Azure Active Directory and Azure Key Vault.
7. Model performance is evaluated using log metrics and monitored using Azure monitor
8. Models are retrained as necessary in Azure Machine Learning.
9. Pre-built simulation and optimization tool available as part of Tiger MMX solution are customized before deploying it on Azure apps for consumption

The business-friendly front end of the solution provides a plethora of pre-built reports & dashboards to consume the findings like

1. Modeling insights (Recommendations for channels, state of market, Channel performance (contributions, Due to charts, ROI, source of change etc.)
2. Recommendations for optimizations (Marginal ROI, Channel response curves, Channel deep dive, optimal allocations & creative mix)

The interactive tool provisioned as part of the solution will help perform what if analysis by simulating marketing budgets and visualizing impact against the base scenario, ability to compare multiple scenarios against each other based on spend, ROI, new relationships etc. There would also be an integrated Optimizer to get recommendations on optimal marketing mix based on defined business constraints