MIM Planning Summary | How does ChannelMix determine the categories for predictions? (Top, Great, Best Match, etc.)

Version 2024.3

Upon running a prediction in your upgraded Plan Performance tab, you are prompted with a scatterplot that might look like the image below. Reading this scatterplot is quite straightforward and can help you make quick decisions, and the guide below will help you learn how!

The Dotted Lines

First, let's look at the dotted red lines in the scatterplot. Where do they come from, and what do they mean?

These red lines are actually references to your target goal (Qualified Leads in this example) and your cost goal that is found in your Goal Pipeline. The Qualified Leads goal will be the horizontal line, and the cost goal the vertical. These lines serve as visual indicators for the data points in your scatterplot; we can quickly identify which prediction will cost more than my goal. Where your predictions fall in relation to both goal lines guides ChannelMix in categorizing your results. 

Picking a Prediction

Great, that makes sense, but I see categories in multiple positions in relation to the goal lines. Why is that? How do I know which one to use?

While the goal lines do guide the categorizations, ChannelMix takes a few other things into consideration:

• Top Prediction:
  • Top Predictions represent the three predictions that are most efficient within three different spending groups (Conservative, Balanced, and Aggressive). Read more about Top Predictions in MIM Planning Summary | How does ChannelMix decide the three Top Predictions?
  • These predictions will be cached and saved for you to use in the Pacing Performance tab. 
  • You may find that these Top Predictions may not best fit your needs. In that case, you can explore the other predictions:

• Great Prediction:
  • Great Predictions are generally reserved for predictions that are found in the upper left quadrant of your scatterplot (above the target goal line, and to the left of the cost goal line).
  • You should start here as these predictions are ideal as the recommendation can result in more of your target for less amount of money! 
  • However, you may find Great Predictions outside of that upper left quadrant. For example, there could be a prediction in the bottom left corner that is also considered a Top Prediction. This would be because it has the lowest Cost Per Qualified Lead of all the other predictions in the scatterplot. This is the most efficient recommendation.
    
    
• Good Prediction:
  • Good Predictions can usually be found in either the lower left quadrant or the upper right quadrant of your scatterplot. 
  • While these are not Top Predictions, these are still advisable predictions because they have a lower Cost Per Qualified Lead than your Goal Pipeline. In this example, this means our Great Predictions have a Cost Per Qualified Lead lower than $119.
  • Use these recommendations if you find yourself having a smaller budget or a larger budget.
    
    
• Maximize Target Prediction:
  • Maximize Target Predictions are reserved for predictions in the upper right quadrant that are not considered a Great Prediction. 
  • These recommendations provide the most amount of Qualified Leads but are less efficient than your historical performance. 
  • These predictions are only recommended if you do not have a budget constraint, or your ultimate goal is to generate the most Qualified Leads as possible. 
  • This prediction provides a larger but less efficient value in relation to your baseline.
    
    
• Not a Recommended Prediction:
  • The remaining predictions are Not a Recommended Prediction. These predictions can appear in two place: the bottom right quadrant or in the bottom left quadrant as we see in the example above. 
  • These predictions achieve your goals but in a very inefficient manner. 

If you already have an established marketing budget, the model can also suggest the best marketing and media mix with a provided budget:

• Your Budget:
  • This prediction returns the most amount of your target with the provided budget. Read more about Your Budget in MIM Planning Summary | Can I set my own budget in my model? 
  • These predictions will be cached and saved for you to use in the Pacing Performance tab. 

If you are in a channel-level dashboard and have selected Cost as your goal, you may have seen an additional prediction category:

• Best Match:
  • This prediction returns the most amount of your target within a margin of the entered cost amount. This margin is indicated by the shaded grey area. 
    
    • By default, the margin is 10% of your cost. 
  • While there may be other predictions that are more efficient or return more of your target, the purpose of this categorization is to indicate which prediction is best considering the amount of budget you have available.  

With your business objective in mind, use the prediction categories to quickly plan your upcoming marketing strategy! Once you have settled on a prediction, click on it to reveal the recommended marketing and media mix!