Nanigans launches incrementality optimization & reporting solution
The new solution focuses on driving incremental revenue growth from Facebook, Instagram, Twitter and programmatic retargeting campaigns.
Nanigans, the cross-channel SaaS (software as a service) platform for large-scale performance advertisers, has launched incrementality optimization and reporting in the platform, which supports Faceboook, Instagram, Twitter and programmatic retargeting campaigns.
The machine learning-driven solution aims to target consumers deemed likely to be influenced by advertising and limit spending on users who are already likely to convert.
Ric Calvillo, Nanigans co-founder and chief executive officer, said in a phone interview that the new service uses machine learning to predict revenue lift from an impression (a user) and make a bid based on that prediction in real time.
In comparing Nanigans Incrementality to a multitouch attribution (MTA) model, Calvillo said, “Giving partial credit to channels is broken. It’s just de-duping conversions but still using touch-based attribution. That confuses correlation with causality.” That’s because MTA gives credit to the impression or click even when those people would have purchased organically anyway without the extra ad exposure. Nanigans measures revenue lift relative to a holdout sample of people.
“The attributed conversion rate could look really good, but,” said Calvillo, “to maximize revenue, you want to show ads based on lift even if the conversion rate is low.” With this kind of optimization strategy, someone who has already clicked on an ad and been added to a retargeting list is not seen as incremental. “The system would actually bid that down. The second or subsequent ad click won’t be worth as much,” explained Calvillo.
The company uses regression modeling based on repeat purchases and how conversion rates change because of ad exposures. “This sidesteps the problem with attribution because you don’t have to assume any kind of attribution or conversion window. The conversion and spend is measured at the time of occurrence,” said Calvillo,
“We’ve seen big boosts in profitable incrementality,” he said, pointing to a test with Rue La La. The e-commerce company tested the incrementality optimization strategy for its site retargeting campaigns this summer, moving away from a last-click strategy. The shift to incrementality meant Rue La La was either no longer bidding at all or bidding at much lower rates on past site visitors that the system predicted would convert without having to see an ad. Focusing on users with higher predicted incremental revenue lift meant the system raised bids or bid for the first time on users that been neglected under the old last-click bidding strategy. According to Nanigans and Rue La La, the e-commerce site saw a 6.5x increase in incremental revenue from the same level of ad spend after shifting to the incrementality strategy.
Calvillo adds that making this shift means “You have to change the way you measure. The legacy system claims credit based on who you touch. This gives credit across the whole audience.”
Each campaign has its own model, and advertisers can see the actual model within the Nanigans platform and get insights on trends that can be applied in other marketing efforts. To use Nanigans Incrementality, there is a minimum site requirement of 1 million monthly unique visitors.
About The Author
How Machine Learning Works, As Explained By Google
Braze launches an app ecosystem with more than 45 integrations
Today's CMO must tackle these 3 challenges
Building an Amazon Alexa Skill is so easy, Grandma can do it