Snapchat taps AI to boost ad targeting
- Snapchat recently added a “goal-based bidding” (GBB) ad tool for marketers, according to a report in Ad Age. The company said 20% of advertisers on its platform are already using GBB, and that it has boosted campaign efficiency.
- GBB allows advertisers to choose performance metrics beyond just raw views when bidding for ad units on the app. One example would be engagement, with the tool monitoring swiping on an ad as a success metric.
- The new bidding also uses machine learning to target engagement-based ads at Snapchat users who are more likely to swipe. This could up the time users spend in Snapchat ads.
Low ad engagement on Snapchat has long dogged marketers, with some formats averaging less than 3 seconds for viewability. The new GBB system could help brands better reach the right users at the right times in what's another sign that Snap Inc. is improving its features and services for marketers in advance of an initial public stock offering expected early in March.
As Ad Age points out, Snap's goals with GBB could stretch beyond simple engagement metrics like swipes to include areas like app installs and e-commerce. The company reportedly acquired the Israeli AR startup Cimagine for up to $40 million in late December. Cimagine has a focus on scanning real-world products for brands like Coca-Cola and Shop Direct, suggesting that retail and e-commerce may indeed be top of mind for Snapchat as it looks to build out its platform to be far bigger than video messaging. Snapchat is also reportedly preparing to launch app install ads this year, an area that has been dominated by Facebook.
Historically, Snapchat has had an almost adversarial relationship with marketers wanting to reach its young-skewing user base, but the app has rolled out a variety of targeting and measurement features over the past 12 months. GBB's use of machine learning to target users more likely to reach specific campaign objectives is another indicator that artificial intelligence is increasingly important in this field, with the potential to automate and optimize ad buying, selling and delivery.
Follow David Kirkpatrick on
Follow Peter Adams on Twitter