Understanding the Insights section in Morpheus (Marketing Mix Modeling)

In this article, we'll explain how to interpret graphs in the Insights section of our Morpheus MMM platform.

The Insights section in Morpheus is designed to help you configure and train the model and understand what has happened so far in your strategy.

Training Model Configuration

To train the model, you need to set several parameters:

1. Select the KPI Column: Choose the key metric for your Marketing Mix Modeling (MMM) analysis, such as Sales, App Registrations, Customers, Ad Impressions, Ad Clicks, Ad Revenue, etc. 

Note that only one KPI can be selected per model.


2. Select the Date Column: Choose the column containing the dates for your dataset.


3. Select Context Variables: These are metrics that can influence your KPI but are not direct investment channels. Examples include seasonal trends, competitor data from Google Trends, weather data, economic indicators (e.g., inflation), promotional and pricing strategies, industry trends, and events.


4. Choose Columns to Ignore: Exclude columns with data that do not affect the result to avoid noise. 

If the uploaded data has more than one KPI, select the ones you have not chosen to analyze to ignore them.


5. Select Training Quality: Determine the duration for model training:

    • Fast (0-3 mins): For immediate results.
    • Performance (3-5 mins): A balance between speed and accuracy.
    • Accurate (6-10 mins): For precise results, requiring more time.


After setting these parameters, click "Train Model" to start the training process, which will take a few minutes depending on the selected training quality.

MMM_icon_c-1



Training Results 

Once training is complete, you will receive the results in the form of visual graphs that we explain below.

1. Model Accuracy and Margin Error

They are displayed in the boxes on the left side, showing the model's accuracy percentage and margin error. 

You can read more about model accuracy and margin error here


2. Response Model Accuracy

This graph shows how well the response model predicts outcomes based on current consumer and market responses to the marketing mix data. It helps you evaluate the model's accuracy by comparing predicted and actual results over time.


3. Current Media Mix

It's a pie chart illustrating the distribution of your current spending across media channels, shown either in percentage or monetary terms.


4. Media Effects - Average

This graph demonstrates how spending on different media channels contributes to your KPI. It shows which channels have the most significant impact.


5. Media Effects - Weekly

This graph shows the impact of each marketing channel on business outcomes, split by weeks. It helps identify fluctuations, pinpoint high-performing channels, and determine optimal campaign timing for informed strategy and budget allocation.


6. Funnel Effect

This graph shows the impact of investments in channels on the left side on the channels at the bottom. It displays how investing in one channel increases or decreases the effectiveness of another. Rows represent upstream ads (investments), and columns represent outcomes. Each cell shows the percentage impact, with positive values indicating an increase and negative values indicating a decrease.


7. KPI Summary Breakdown

This graph shows how each channel affects the KPI over time, with weekly or monthly views. Values that negatively impact performance are highlighted in red, while those that contribute more to the KPI are in blue. Different shades of blue and red indicate the intensity of impact, making it easier to spot trends.


8. Lag and Carryover

This graph shows how advertising impact varies over time. The carryover effect shows how the effectiveness of an advertising campaign diminishes over time. The lag effect is the delay between advertising and its response. This helps understand when an advertising channel starts contributing to your KPI, how long the effect lasts, and when it diminishes. You can also select specific channels to display by clicking Hide and Invert


9. Ad to Consumer Response

This graph shows how adstock deviates from advertising spend over time. It demonstrates the impact of the campaign on the audience after the advertisement stops, showing diminishing returns and the lag and carryover effects by year or week.


10. Non-Media Contribution

This graph shows how non-media factors, which are not related to advertising spending, influence the KPI. It displays both positive and negative impacts, with the option to hide and show different factors by clicking Hide or Invert


11. Shape Effect (Diminishing Returns)

This graph shows how sales change with increased advertising intensity within the same period. It illustrates the saturation curve or diminishing returns curve, helping understand the optimal spending per media channel for the best profitability and when saturation begins.

These insights help you understand the detailed influence of different variables on your KPI, aiding in further optimization and planning. For additional questions or support, contact us via website chat or email. We're here to help!

Request access to Morpheus MMM.