Channel Spend Optimization & Prediction (Media Mix Model)

With the ChannelMix spend optimization and ROI prediction dashboard, you can 1) calculate the optimal marketing mix based on your available budget and 2) understand the impact of shifting spend to different channels in your marketing mix. 

 

This dashboard is powered by the ChannelMix modeling engine, which puts 2+ years of historical performance data into a machine learning model to calculate the best possible outcome whether that’s leads, sales or some other target. 

To use the Spend Optimizer, first input your budget in the top right corner. In this example, we have an available budget of $1,000,000. Then hit the Run Optimizer button. Typically, the Optimizer takes about 15 minutes to run through all scenarios to find the one with the highest ROI. 

marketing_spend_optimization_-_budget_input.png

 

With a budget of $1,000,000, the Optimizer is telling you to spend $90,780.58 in Display, $147,813.16 in Paid Search, $154,248.67 in Paid Social and so on. This marketing mix will result in 3,256 leads generated with an ROI of $31.56. 

optimizer_results.png

 

For reference, the dashboard will show you historical results for every time you’ve previously run the model.

marketing_spend_optimization_-_historical_results.png

 

Now let’s say you’ve already spent your TV ad budget for the rest of the year, but the Optimizer told you to spend $506,839.01 on TV. This is where the Predictor comes in. Let’s see what happens when we shift the TV budget to Paid Search. 

In the Predictor, we’ll input spend for each channel using the values from the Optimizer — except for TV. We’ll add that spend to the recommended Paid Search spend. Now we’re spending $654,652 in Paid Search and $0 in TV. 

The Predictor shows that shifting money from TV into Paid Search will result in fewer leads and a lower ROI. 

predictor_results.png

 

And that’s how you can use the ChannelMix modeling engine to optimize your marketing mix and predict ROI. 

Was this article helpful?
0 out of 0 found this helpful
Have more questions? Submit a request

Comments

0 comments

Please sign in to leave a comment.