Previously, we’ve shown that our “likely to buy” machine learning model can improve ROAS by 2x. But exactly how big is the difference between “likely to buy” visitors and “unlikely to buy” visitors?
Thanks to one of our clients, we have an answer.
Figleaves, a multi-national apparel brand, activated their machine learning audiences on their Google display campaigns to learn exactly how much money they were wasting on non-converting visitors.
In a test running for about two weeks, the “likely buyers” audience yielded 6.9x ROAS compared to the “unlikely buyers” audience.
Not only that, but the likely buyers also yielded 2.5x revenue on only one-third of the ad spend compared to the unlikely buyers.
And just like that, we know exactly “which 50%” is being wasted. To us that looks like pretty conclusive reasoning to cut ad spending to the unlikely buyers and save a ton of money.
Get in touch with us to find out how much your brand can earn back from your advertising with our solutions.
Share this!
