A new Facebook feature allows publishers to reach audiences who are specifically interested in such narrow subjects as “computer servers” and “digital data” with their posts without alienating the ones who might enjoy reading about broader topics like “technology” or “consumer electronics.”
Released last month, Audience Optimization replaces the company’s earlier Interest Targeting feature, which prevented people who didn’t match the interests assigned to a post from seeing it in their news feed.
The newly developed tool includes three key features, Facebook Inc.’s product manager Peter Roybal wrote in a Jan. 21 blog post: preferred audience, which allows publishers to add interest tags to their posts that users with matching interests will have the highest likelihood of seeing; audience restrictions, which allows publishers to specify the audiences that would not find a post relevant, based on such factors as location, language, age, or gender; and audience insights, which allows publishers to monitor how well their content performs, complete with reach and engagement measurements for each interest tag.
Most importantly, the Audience Optimization tool does not prevent posts from appearing anywhere else on Facebook, unless a publisher specifies otherwise. In an FAQ, the company recommends attaching between six and 10 interest tags to each post, with a maximum of 16 available. During tests, the company said that organic reach remained more or less the same once the tags were added, while audience engagement went up.
It’s worth noting that while Facebook provides publishers with literally hundreds of thousands of interest tags, which are drawn from the company’s extensive data sets including public posts, Facebook ad tags, shared pages, and interests, advertisers cannot create tags of their own.
How does it work?
For those interested in learning how the tags were compiled, a fascinating article at the Verge illuminates the process, a still-imperfect algorithm which has determined that, for example, “MS-DOS” has a larger audience than “PlayStation” (likely because the former is made up of two common letter combinations, the Verge notes), and that the “comitative case,” a grammar term for words like “with,” apparently has an audience of 58 million (likely because “with” is a popular word, the site says). For those concerned the interests assigned to them reflect an invasion of privacy, a Facebook spokesperson told the Verge that its Messenger service was excluded from any analytics.
By emulating a search that entered all possible letters and numbers until no other results were returned, the Verge also created a list of Audience Optimization’s 2001 most popular tags, ranking them according to the number of users associated with each.
Some of the more narrow tech-related interest tags include “cloud computing” (associated with more than 235 million accounts), “tablet computers” (associated with more than 200 million), “personal digital assistant” (over 85 million), “computer monitors” (over 52 million), “web search engine” and “BlackBerry” (both over 49 million), and “phablet” (over 46 million).
How to use it
As for how marketers can best take advantage of the new service, Facebook conducted case studies with several media companies, including MTV.com and the New York Times, and came back with a few words of wisdom.
To start, the company suggests avoiding the most general categories like “Entertainment” and “Sports,” since someone interested in movies might not watch reality TV, or someone interested in the NFL might not care about the NBA.
For stories about public figures such as Lady Gaga or Justin Trudeau, Facebook saw a direct correlation between an article’s subject and audience interest, making direct tags most effective when well-known names – or brands – were involved (Google Play and YouTube are both high on Verge’s list). Indirect tags, meanwhile, turned out to be most effective in connecting the news with people who were interested in the subject being covered, and would therefore consider an event’s larger significance.
For example, in a story about U.S. President Barack Obama’s final State of the Union address, the New York Times used “Barack Obama” as a tag to target fans of the incumbent president, “Democratic Party” to target his supporters, “Republican Party” to target his opponents, and “United States Congress” to reach people interested in the topic in general – resulting in an increased number of clicks and likes, Facebook said.
Presently, Audience Optimization can only be used to add interest tags from Facebook’s desktop website, and the company does not yet have any plans in place to support tags in languages other than English. You can find out more about the feature here.