Once you've built your model, the next step is to evaluate the model fit, interpret the results, and use the results to make better marketing decisions. This post-modeling phase is where you'll uncover key insights about your marketing effectiveness and plan for the future. This section will guide you through understanding your model's outputs, debugging common issues, and using the budget optimizer.
| Page | Description |
|---|---|
| Assess the Model Fit and Results | How well does your model actually match your data? This page explains key metrics like R-squared to evaluate your model's "fit" (or "goodness-of-fit"). You'll learn how to assess your model's predictive accuracy and understand the limitations of these metrics for causal inference. |
| Incremental Outcome, ROI, mROI, and Response Curves | This page explains some of the most important outputs of your model: the response curves, ROI, and mROI. You'll learn how to read response curves to understand the diminishing returns of your marketing spend, and the difference between ROI (the overall return) and mROI (the return on your next dollar spent). |
| Assess the Baseline | This page explains what the "baseline" in your model represents. It's the expected outcome if you hadn't run any of your marketing activities. Understanding your baseline is key to measuring the incremental impact of your marketing efforts. |
| Interpreting Visualizations | Meridian produces a variety of plots and charts to help you understand your model's results. This guide walks you through the key visualizations, explaining what each one shows and how to use them to assess your model's performance and gain insights into your marketing activity. |
| Interpreting Optimizations | After running a budget optimization, this page helps you understand the results. It explains the different outputs of the optimizer, like the recommended budget shifts and the expected gains in your KPI. You'll learn how to interpret these recommendations to make informed decisions about your future marketing budgets. |
| Refreshing the Model | Your marketing landscape is always changing, so your model should too. This page provides best practices for updating your model with new data. It covers how often you should refresh your model and the steps to take to ensure it remains accurate and relevant over time. |
| Model Debugging | This page discusses common issues that can arise during the modeling process and how to address them. From unexpected ROI values to problems with model convergence, this guide offers practical tips for troubleshooting and improving your model. |
| Optimization with Reach and Frequency | For channels where you have reach and frequency data, the optimization process is more nuanced. This page explains how Meridian's budget optimizer uses this richer data to provide recommendations not just on how much to spend, but also on the optimal frequency to aim for. |
| Optimization without Reach and Frequency | This guide explains how to use Meridian's budget optimizer for channels without reach and frequency data. It details the process of finding the optimal budget allocation based on the response curves generated from this type of data. |