Every day, you are spending money on ads. A platform like Facebook will tell you how many impressions you’ve got. The more your ads are printed, the more people you are reaching. Some might click on your ad, and some will convert into whatever you want them to convert to: app users, gamers, or paid customers.
You might think there is a direct correlation between impressions and people reached. Or people reached and clicks. After all, the more people you reach, the more clicks you get.
Until you create a correlation matrix. How strong is the correlation between impressions and clicks? If it is weak, your campaign might not be performing well.
Take for example reach and frequency. Yes, you are reaching more people every day, but over time you might be reaching one person only twice a week. That is a very weak frequency if you want them to remember your brand.
If you go to Facebook Ads Reporting, do a breakdown by day, and choose ‘Clicks (All)’, ‘Amount Spent (USD)’, ‘Reach’, ‘Results’, ‘Impressions’, ‘Frequency’, export the data to CSV as covarience.csv, you can run this script and get a correlation matrix like this one:
- ML Models for Threat Prediction in Ads - 04/23/24
- The Science Behind Google AdMob’s Less Intrusive Ads in Mobile Games - 04/22/24
- Out-of-Home Advertising Meets AI - 04/15/24