On March 19, Facebook settled a civil rights lawsuit with the National Fair Housing Alliance (NFHA), Communications Workers of America (CWA), several regional fair housing organizations, and individual consumers and job seekers.
The resulting agreement includes a long list of changes to prevent discrimination in housing, employment, and credit advertising across Facebook’s suite of advertising platforms.
Facebook agreed to make eight significant changes including establishing a separate advertising portal for ads in the above verticals that will including limited targeting options and creating a lookalike ‘lite’ tool called ‘Similar Audiences.’ This is in addition to removing age, gender, and Zip code targeting.
But what does this really mean for advertisers, end consumers, and the wider digital advertising eco-system? The benefits of targeting and personalization are clear, yet used unscrupulously, can be discriminatory.
Breadth and waste
It’s interesting to note Facebook is taking initiative, even calling themselves ‘pacesetters’ – a stark contrast to previous moves. And a lesson learned from a difficult 2018 with the Cambridge Analytica scandal still reverberating, while Brexit and 2020 US elections loom not too far in the background.
I can tell you firsthand that over the past 18 months the party line from Facebook to agencies is, “Go broad and trust our algorithm”; i.e. target everyone (220m users in the US) and let Facebook’s algorithm find your highest-converting audience. The cynic would say that’s a great approach to drive more media spend and, ultimately, revenue for Facebook.
This poses a challenge, as budgets need to be sufficient to allow algorithmic learning, whilst agencies need to continually improve the perceived value of their media buyers.
Through our own multi-variant tests, it’s clear the algorithm works well – even though Facebook’s recently launched campaign budget optimization (CBO) is seemingly slower than its adtech counterparts. Instead, though, by adopting a strategic approach to segments and audiences through insightful research, we can expedite performance.
Consider: In one test we developed a data model to allow us to seamlessly track digital media through to true offline conversions to one client’s fully funded loan product. This allowed us to optimize towards loan origination value rather than website form fills, leads or clicks. After all, if you’re a financial services company, for example, why target ‘18- to 24-year-olds who rent’ for home equity loans? The only reason: now, when advertising on Facebook, the regulatory agreement determines we have to target everyone. If we were to adopt Facebook’s broad approach based on its legal settlement, we estimated 7-12% of media would be wasted in upper funnel media.
Not only is this media wasted from a client perspective in media spend, ad-tech, and agency fees, it also delivers a poor user experience. The mantra of digital advertising has always been relevancy. Once this is stripped away from an advertiser’s toolkit, then confidence, and performance, in the channel could be diminished. Rather than right time, right place, right message, it’s more all the time, all the place(ments), and hopefully) the right message – if it’s even the right audience.
As we await the true implications of these changes — not only with Facebook but with other digital media channels that rely heavily on demographic, Zip code, and recruitment information — one can’t help but ponder if this could be a damaging blow to LinkedIn’s advertising offering.
‘Targeting’ without discrimination
In the meantime, it’s possible to connect with target customers in other ways.
Marketers can use built-for-purpose creative customized to each Facebook placement with ad copy that allows audiences to self-qualify. Often marketers over-index their post copy efforts on search, for obvious reasons. But A/B testing post copy in social while leveraging behavioral science effectively ‘nudges’ consumers to identify if they’d pre-qualify – and in turn drives higher conversion rates.
Another option is developing a Messenger strategy. We are seeing huge growth and success in driving audiences through to Messenger, where a chatbot can qualify a lead in a seamless fashion and automate the sales process from ad through to CRM integration and even direct to specialist sales teams.
Most important is to adopt a multi-layered approach to audience targeting. We predict the continuing removal of ‘in-market’ data, which comes at a premium despite diminishing impact. So, avoid over-indexing on third-party data and prospect smartly through affinity-based audiences. Balance the weighting of media toward intent-based signals. And track conversions that deliver meaningful business outcomes; upper funnel optimization events can be misleading, eventually leading to sub-optimal decisions.
Fair Lending is just one area where new regulations are changing how marketers can reach out to consumers. We are now in a highly dynamic world where marketing strategies need to be adaptable. Focusing on media, channels or audiences with a myopic lens exposes significant business risk. This is the joy and pain of operating in the paid media world: one minute you’ve mastered the art of delivering value, and then, faster than you can say Cambridge Analytica, the world changes once again.
More and more, marketers are starting to deliver multichannel media strategies that focus on long-term benefits rather than short-term gains and, ultimately, put the audience at the heart of the conversion journey.
This can only be a positive evolution no matter what regulatory, or media platform, changes we face along the way.
Greg Allum is vice-president, display and social, at LQ Digital