Enrich your marketing mix model with AI-powered macroeconomic data from the Federal Reserve Economic Data (FRED) database. By incorporating external context, you can isolate the true impact of your marketing from broader economic trends.
Prerequisites
- Existing Project: You must have at least one data source with a dependent variable (your business outcome, like revenue) so the system can calculate correlations.
- Project Context: Ensure fields like country, industry, and product categories are completed to help the AI suggest the most relevant series.
Step-by-Step
1. Open the FRED Discovery Dialog
- Navigate to your project’s Data tab.
- Select Add Data → Discover FRED Indicators.
- The FRED Discovery Dialog opens, acting as your navigation panel for economic insights.
2. Review AI Suggestions
The platform scans your project metadata to propose relevant economic series. For each suggestion, the system provides:
| Column | Description |
| Series Name | The official FRED title (e.g., “Consumer Price Index”). |
| Description | A brief explanation of the economic metric. |
| Correlation | A score showing how closely the series moves with your sales. |
| Frequency | How often the data is updated (daily, weekly, monthly). |
The results are ranked by correlation strength, placing the most statistically relevant drivers at the top.
3. Select and Import
- Check the box for each indicator that plausibly influences your business, such as Consumer Sentiment or Unemployment Rates.
- Click Import Selected to add these as a new data source card.
After Import
- Contextual Mapping: These fields are typically assigned as Context variables during run configuration.
- External Regressors: The modeling engine uses this data to account for external swings in demand, preventing the model from misattributing market growth to your media spend.
Common FRED Indicators for MMM
| Category | Example Indicators |
| Consumer Sentiment | Consumer Confidence Index, Michigan Sentiment Index. |
| Price Levels | Consumer Price Index (CPI), Producer Price Index (PPI). |
| Employment | Unemployment Rate, Initial Jobless Claims. |
| Interest Rates | Federal Funds Rate, 10-Year Treasury Yield. |
| Retail & Housing | Retail Sales, Housing Starts, Building Permits. |
Tips for the Pilot
- Date Alignment: FRED data is often monthly; the system automatically interpolates it to match your project’s daily or weekly timeline.
- Avoid Overcrowding: Adding too many variables can lead to “overfitting”. Start with 1–3 highly correlated indicators to maintain a clear roadmap.
- Iterative Discovery: As you add more sales data, re-run discovery to see if new economic patterns emerge.