To effectively converge Multi-Touch Attribution (MTA) and Marketing Mix Modeling (MMM), it is essential to integrate the strengths of both approaches, aligning their methodologies and insights for a cohesive and unified measurement strategy.
Here is a step-by-step guide to achieving this convergence:
Step 1: Align business goals and key performance indicators
Set clear, unified goals: Start by clearly defining the business goals and the key performance indicators (KPIs) that the MTA and MMM will measure. Make sure that both models are aligned on what success means to the business. Typically, the choice will be conversion value (e.g., revenue) or conversion volume (e.g., number of sales).
Step 2: Use Testing as a Gateway
Incrementality Effect Studies: Implement incrementality tests (e.g., randomized control trials, uplift tests) to validate and calibrate MTA and MMM models (see Chapter 2: Validate the choice of model). These tests provide baseline data that can help align the results of the two models, ensuring that attribution is correct and that media mix decisions are based on true incremental impact.
Step 3: Synchronize data entries
Harmonize data sources: Ensure that MTA and MMM use harmonized data, including consistent time periods, audience segments, and touchpoint data. This includes aligning online and offline data sources so that both models operate from the same core data set.
Granular or Aggregated Data: MTA uses more granular user-level data (e.g. clicks, impressions), while MMM typically uses aggregated data (e.g. spend, reach). Ensure that data aggregation rules for MMM align with the touchpoint data used in MTA to minimize discrepancies.
Step 4: Integrate the results for cross-validation
Cross-validation of results: Compare MTA and MMM results regularly. Use MTA to validate short-term channel-specific information, while MMM should validate broader, cross-channel, and longer-term effects. Discrepancies between the two should be investigated and adjustments made if necessary.
Feedback loops: Establish feedback loops in which MTA data is used to refine MMM data (such as channel effectiveness) and vice versa. For example, the MTA can provide information on the relative importance of digital channels, which can be incorporated into the MMM.
Step 5: Adjust MTA attribution models based on information provided by MMM
Adjust credit allocation: Use the information provided by MMM to adjust the credit allocation in the MTA model. For example, if MMM indicates that a particular channel (such as TV) is driving a significant increase in overall conversions, this can help adjust the weight of TV within the MTA to ensure it properly reflects the influence of upper-funnel activities.
Eulerian offers a customizable attribution engine to integrate this new information! Contact your CSM team to learn more.
Step 6: Calibration
Continuous Calibration: Regularly calibrate MTA and MMM models using new data and incrementality testing to ensure they remain aligned with actual marketing activity performance.
Scenario testing: Perform scenario analyses in which the two models are fitted and tested under various assumptions to identify optimal points of convergence.
Step 7: Unified reporting and decision-making framework
Integrated Dashboards: Develop dashboards that present MTA and MMM results side-by-side, providing a unified view that stakeholders can use to make informed decisions.
Consistent decision-making frameworks: Establish a consistent decision-making framework in which the results of both models guide tactical (MTA) and strategic (MMM) actions. For example, use the MTA for in-campaign optimizations and the MMM for budget reallocations in future planning cycles.
Eulerian offers MTA and MMM based performance reports to make these comparisons easier!
Eulerian has also developed a feature for automatic MMM calibration by the MTA. Contact your CSM team to learn more.