The Power Of AI And Performance Of Native Advertising

• Goal: GoBear was looking to augment its native advertising campaigns to further drive conversion of users looking to purchase travel insurance.

• Solution: A highly targeted native advertising campaign that leveraged Verizon Media’s state-of-the-art machine learning system, AdLearn, and Predictive Audiences, with advanced predictive performance algorithms.

• Results: GoBear achieved stellar campaign results through the use of Verizon Media’s Predictive Audiences. GoBear began leveraging Predictive Audiences for their campaign in May, after the traditional peak travel insurance purchase period and still saw significant impact. Cost per acquisition (CPA) reduced significantly — up to 4 times lower than before the campaign launch. Conversions remained healthy throughout. 

Working with Verizon Media, GoBear – Asia’s leading financial supermarket and comparison site – has been enjoying significant success with native advertising campaigns across Yahoo’s platforms since 2017. 

On the back of this successful partnership, GoBear wanted to take its ongoing campaign to the next level through the power of Verizon Media’s cutting-edge artificial intelligence and machine learning capabilities. 

The company aimed to lower CPA whilst still effectively linking their customers with suitable travel insurance providers. Successful acquisitions were measured as users landing on the insurance page, filling out their details and hitting ‘go to provider’.

The right conversations with the right people

Verizon Media’s state-of-the-art machine learning system, AdLearn, which combines extensive demand and supply data with the most advanced predictive performance algorithms in the form of Predictive Audiences, allowed GoBear to reach people more likely to convert. 

The GoBear team wanted to test the effectiveness of Predictive Audiences when utilized alongside traditional site retargeting (where people have previously visited but did not convert) and lookalike targeting (where data on existing users is used to target consumers with similar interests or from a similar demographic).

GoBear leveraged Predictive Audiences to analyze existing converted users to find the most accurate predictors of purchase, before modelling them against Verizon Media’s users to create a series of audience segments. In all, there were eight segments, ranging from tier 1, which had very high probability of conversion, but very narrow audience, to tier 8, where the probability of conversion was a lot lower but delivered a much broader base. The machine learning then intelligently decided on the best mix to target for this campaign.

GoBear’s Campaign Took off with AI 

Predictive Audiences enabled GoBear to exceed its campaign goals, driving a healthy volume of conversions and a higher conversion rate, while delivering a low cost per acquisition (CPA), even without Run of Network advertising. 

Just one month after deploying Predictive Audiences, it began delivering conversions on par with site retargeting and lookalike targeting. 

Crucially, Predictive Audiences continued to drive very low CPAs; boasting CPAs four times lower than prior to the activation of Predictive Audiences during the campaign, even after the peak travel insurance purchase period. 

The campaign’s findings backed Verizon Media’s own internal research which has found that machine learning can help deliver a 48 per cent uplift in conversions, a 38 per cent reduction in cost per click (CPC) and a four to eight-fold reduction in CPA.

Arshant Nair, Regional Performance Marketing Specialist at GoBear: “Verizon Media’s Predictive Audiences tool allowed us to better understand our audiences and to have more relevant interactions with them through the use of machine-learning optimization and forecasting. The ability to reach our customers at scale meant we enjoyed an overall lower cost per acquisition. The campaign results exceeded our expectations. We remain excited about the possibilities to build on this success by partnering with Verizon Media.”

Key Learnings

• Predictive Audiences enabled GoBear to achieve its CPA target, while keeping the volume of conversions very healthy, even without RON advertising.

• Verizon Media’s machine learning and artificial intelligence-led solutions were able to accurately estimate the users likely to convert in a highly cost-effective manner.

• The close, long-standing partnership between Verizon Media and GoBear proved advantageous for the campaign. For Predictive Audiences to work best, clients must be open to implementing tracking on their pages to gather the required volume and quality of audience data.

• The unique combination of mobile-first data, performance boosted by machine learning, and exclusive, high-quality inventory helped GoBear exceed its campaign goals.