Through new partnerships with top online and offline purchase data providers Datalogix, Epsilon, Acxiom, and BlueKai, Facebook is now allowing advertisers to target hashed lists of existing and potential customers, and categories like role-playing gamers or soda drinkers. This expansion of Facebook’s Custom Audiences program could rake in revenue and attract businesses by matching ads to real spenders.
AdAge first reported the deal with Datalogix, Epsilon and Acxiom last week, though now these partnerships are confirmed along with one with BlueKai. It’s part of Facebook’s ongoing quest to show businesses that ads on the social network are not just scattershot brand marketing jazz, but can actually reach the exact customers that buy their products.
That’s been a bit of a struggle, as a combination of complicated products, targeting restrictions, and advertiser inexperience has led some businesses to fail to produce results. These businesses sometimes go out of their way to blast Facebook’s ad platform as ineffective, scaring away other big advertisers.
This is partly why Facebook launched the Custom Audiences program in September. It’s an important complement to Facebook Exchange, which allows for real-time, cookie-based retargeting of people who visit an advertiser’s website. Data providers can use cookies to add people to Custom Audiences, but can also use email sign-ups and other contact information.
The four new partners already work with many of the world’s biggest brands to target email marketing, direct mail, and other online ads. Now they’ll be able to use the same ones to pinpoint Facebook ads, as Facebook explains on its Studio Blog. Businesses can request their data partner upload a privacy-protected, hashed list of customer or potential customer information such as email addresses. The advertiser can then target these sets of people, but don’t ever know the identity of who they’re reaching.
Businesses who don’t already work with one of these data providers can still take advantage of the partnership. Facebook is setting up pre-defined categories like auto-intenders (people planning to buy cars), luxury fashion-buyers, and more that any business can target.
Anecdotal tests of the program provided by Facebook show promise, with Chicago car dealership Castle Auto Group seeing a 24X return on their ad spend through a combination of Facebook offers targeted to Custom Audiences of their existing customers. Hong Kong game developer Klingnet reduced their cost per install by 40 percent by using Custom Audiences of people likely to enjoy their titles.
Along with data protections and hashing to keep advertisers from determining the names or identities of people located by the third-party data providers, Facebook is offering users opt-outs from the program. Facebook went so far as to publish a Note on its Privacy Page detailing how people can click on Custom Audiences ads to see what specific advertiser was targeting them, hide future ads from that advertiser, and be exempt from that data provider’s targeting.
If the Custom Audiences program works, a Honda dealership could go from targeting “people who like Honda in Chicago” to “people in Chicago shown to have visited car-buying websites, signed up for dealership email, or bought a Honda between three and ten years ago.” As you can imagine, the latter is a lot more likely to lead to a sale. If businesses make money with Facebook ads, they’ll spend money on Facebook ads.
This is all part of a grander shift of online ads getting more accurate. With time, more offline and online purchases and behavior are going to be tied to demographic clusters and individuals. Out will go the vaguely targeted shotgun approach ads of TV and print. In will come laser-pinpointed ads that reach people genuinely ready to spend money on a product. This bodes very well for online ad platforms where people spend a lot of time including Twitter and media sites, but most of all Facebook. If you have a captive, always-there audience, you’re going to make a fortune once measurement and targeting tech evolve.
Comment