Facebook Marketing Science and Oracle Data Cloud (ODC) conducted a meta-analysis of 64 US automotive campaigns that ran between February 2015 and March 2016 across a variety of publishers.
Facebook’s Auto Measurement Lead, Brandon Herring, explained, “We conducted this meta-analysis to demonstrate the tradeoffs between audience size and quality within the Auto market.”
Our top finding is that auto advertisers should use CRM data and website converters to build a high-quality seed audience and then use lookalike audiences to find more people like them to balance scale and quality.
The ODC analysis first looked at the publishers of the 64 US automotive campaigns, evaluating each publisher’s performance against two metrics:
We also worked with ODC to evaluate how well Facebook performed relative to the top-performing publisher for each metric in each study.
There were three primary metrics in this meta-analysis:
Audience outcomes were measured by ODC’s Buy-Through Rate (BTR) methodology, combining impression data from Facebook and other publishers with registration data from IHS Automotive (driven by Polk) to measure the concentration and number of car buyers within a target audience.
We also looked at our internal data on top of the analysis with ODC and found that lookalike audiences presented the strongest targeting opportunity on Facebook.
Across all of the publishers measured, we saw that niche auto-buyer websites typically had highly curated audiences and provided a higher percentage of US car buyers—but this was typically at a lower scale than can be found on Facebook. Advertisers were able to reach more US car buyers on Facebook than through most other publishers. In this analysis, Facebook’s audience was above average on the quality index and was particularly strong in delivering auto buyers at scale, ranking in the top 10% of publishers in these studies and first on 25% of the campaigns studied.
Combining quality and scale resulted in the best performance for Facebook compared to the other publishers in the analysis. Facebook was in the top 10% of publishers for the combination of scale and quality, resulting in highly efficient reach for auto advertisers. Facebook was also ranked first in 15% of the campaigns studied when combining quality and scale.
In addition to analyzing overall performance, we evaluated the results of 48 of the US automotive campaigns based on their targeting strategy on Facebook. Campaigns were categorized on their primary targeting strategy on Facebook: Core Audiences, Custom Audiences and Lookalike Audiences.
Campaign performance varied. Facebook’s Custom Audiences provided the highest concentration of quality audiences on Facebook but at a limited scale. Interest-based targeting within Core Audiences provided the highest scale but with lower quality for reaching car-buying audiences. We found that Facebook’s Lookalike Audiences was able to deliver a balance of the two metrics with a high-quality audience at scale.
Reviewing results from an ODC-measured US automotive campaign, Jenny Reed, VP Media Director of 22squared, explained that “using the Oracle Data Cloud data, paired with Facebook and Instagram’s scale, Southeast Toyota and 22squared pinpointed tactics that influenced real shopping behavior. The ability to do that at scale has now changed the way we look at both our social and tactical strategies.”
Facebook recommends high-quality focused audiences for lower-funnel direct response campaigns with conversion objectives. Quantity is more important for upper-funnel brand campaigns with reach objectives.
Prioritize what reach is needed to deliver your campaign goals when identifying the right publishing partner. Niche sites may have a high percentage of US auto buyers but cannot always provide sufficient scale needed for larger reach campaigns.
In this analysis, we saw that the best-performing campaigns on Facebook used Lookalike Audiences to maximize both quality and scale. Seed those audiences to build a high-quality target by using CRM data and website converters.
As we’ve seen ring true for other brand marketers, reaching more people can overall be a more efficient strategy a more efficient strategy. For example, Facebook saw the strongest performance when US automotive advertisers targeted qualified audiences at scale. Changing attitudes and behaviors at scale will ultimately lead to growing business opportunities.