Following a recent uptick in data-driven ad offerings and a rapidly expanding ad exchange, Facebook is responding to fresh privacy concerns with new details about how it collects user data and how users can opt out of certain features. A series of mechanisms have been put in place to protect user privacy under the framework of Facebook Exchange (FBX), custom audience targeting and Facebook’s now year-long partnership with data matching firm DataLogix.
Facebook and DataLogix also released a study today, highlighting the dichotomy between Facebook’s need to prove the value of its platform to advertisers while convincing consumers and privacy advocates that it protects user privacy.
Almost four months have passed since Facebook launched its real-time ad exchange and it continues to grow the exchange with new partners, most recently Criteo and Rocket Fuel. However, with FBX enabling Facebook to connect its platform to third-party data for targeting ads, it was only a matter of time before privacy issues bubbled to the surface.
Facebook works with third-party service providers to agree on a unique ID number for each visitor’s browser, explained Joey Tyson, privacy engineer at Facebook, in a blog post today.
“If someone then visits Facebook and his or her browser has that ID, we notify the service provider, who tells us when a marketer wants to show a particular ad. This allows marketers to show you ads relevant to your existing relationship with them – and without them needing to send us any personal information about you,” noted Tyson. “When we show an FBX ad on Facebook, it includes an “X” link that lets you provide feedback about ads. We also provide a link that lets you learn more about and choose to opt out of future ads from the service provider responsible for that ad.”
Although Tyson made it clear that the individual IDs for use on the ad exchange will be separate from users’ Facebook IDs, he did not provide detail as to whether the strings of code would remain interconnected. In other words, there could be a link between IDs associated with a user’s browser and his Facebook account ID.
Tyson also provided more detail on custom audiences, a new feature that enables marketers to reach specific Facebook users like previous customers, for example. He explained that brands can provide “hashes” of their customers’ email address to match their ad with those customers on Facebook. “These hashes are bits of text that uniquely identify a piece of data (such as an email address) but are designed to protect against reverse engineering which would reveal that data,” he wrote. “If a hash from the store does not match any of ours, we discard it without ever discovering the corresponding email address and without storing any information that we did not have before. And once we no longer need the hashes that do match, we delete them too.”
After Facebook’s partnership with DataLogix was uncovered late last month, it has generated scrutiny from privacy advocates even though Facebook does not share personally identifiable information with the firm, which measures the impact of Facebook ads on in-store sales. Although Facebook has already stressed its commitment to ensuring the privacy of its users through the effort, Tyson readdressed the issue to allay concerns.
He described the goal of Facebook’s partnership with DataLogix as an attempt to replicate “similar studies” that companies are accustomed to, referencing traditional media such as newspapers, TV and radio. Facebook compares hashes of some Facebook data with hashes provided by DataLogix, Tyson explained. “Once we compare, we are able to send corresponding data on the reach of large-scale ad campaigns, which Datalogix uses to create aggregate reports comparing product purchases by large groups of people who did or did not see an ad.”
He also noted that just as DataLogix is not privy to Facebook profile info, Facebook does not receive any purchase data from the firm. “We cannot specifically tell whether or not you purchase a marketer’s product,” Tyson continued. “DataLogix only sends the marketer aggregate information about large groups of people. None of this data is attributable to an individual Facebook user.”
Coincidentally, a study released by both firms today at the IAB MIXX conference highlights the fact that Facebook must balance privacy with the need to prove the value of its platform to advertisers. After studying results from more than 50 digital campaigns, Facebook drew three conclusions: impressions create value, reach drives revenue and message frequency is key.
The firms found that 99 percent of sales generated from online branding ad campaigns on Facebook were from people who saw an ad but did not click through. The study also showed campaigns with maximum reach averaged a 70 percent greater return on investment, according to Brad Smallwood, head of measurements and insights at Facebook.
He added that when brands reallocated high frequency impressions to people seeing too few impressions, they typically experienced a 40 percent increase in ROI with the same budget. “While these conclusions might seem familiar to traditional marketers who use TV, they represent a substantial shift from the focus on click-optimization that is more typical of digital campaign planning,” Smallwood wrote in a blog post about the study.
President Trump's digital savvy isn't limited to social media. As it turns out, the Trump Organization owns thousands of domain names, possibly even more than 10,000.
Silicon Valley loves fancy job titles. It’s just something we do, and software and technology lend themselves to it. But it’s not always helpful.
Social media has developed into an effective component of digital strategy, but measuring its performance is still a challenge. How will analytics affect social media in 2017?
I didn’t vote for him last November. There was no way this registered Democrat from the blue state of Massachusetts would check that box. But I have to give him props for his tweets.