Labels on ads with political content in the Faccebook News Feed (Source: Facebook webinar)

What small publishers need to know about Facebook’s policy on ads with political content

Facebook announced on May 24 that any Facebook Page promoting political content on the platform will need to follow the platform’s rules for “Ads with political content,” including a prominent disclosure of the source of funding and completion of a cumbersome authorization process. Such ads will also be archived for up to seven years in a publicly searchable database. The new policy aims to bring transparency to ads with political content in an effort to preserve election integrity on the platform.

Facebook’s definition of “political content” is broad, potentially sweeping up any promoted news and information on the platform that relates to civic matters. Publishers writing about candidates or “issues of national importance” will find their boosted posts or advertisements rejected if the publisher has not followed Facebook’s new guidelines.

To shine a light on the process, Facebook’s Julia Smekalina and Varun Shetty (News Partnerships) held a webinar with publishers on May 30, during which they answered questions directly from representatives from the media. I attended on behalf of the Center for Cooperative Media, which coordinates with groups including the Local Independent Online News Publishers and the Institute for Nonprofit News. The following are my takeaways for news publishers, beginning with Facebook’s definition of political content and my understanding of how content is being reviewed.

The flow to the archive. Source: Facebook webinar

What are ‘Ads with political content’?

Facebook’s definition of “ads with political content” — which it distinguishes from “political ads” — is perhaps the most contentious part of the new policy.

Here’s how Facebook defines it on its help page:

Ads that include political content are defined as any ad that:
- Is made by, on behalf of, or about a current or former candidate for public office, a political party, a political action committee, or advocates for the outcome of an election to public office; or
- Relates to any election, referendum, or ballot initiative, including “get out the vote” or election information campaigns; or
- Relates to any national legislative issue of public importance in any place where the ad is being run; or
- Is regulated as political advertising.

In practice, the definition is astonishingly broad and may affect any promoted content that relates even vaguely to the political system — including any post mentioning an elected official (or even the title of one), terms of democratic process or governance, or variations of terms related to “national legislative issues of public importance” such as “budget,” “education,” or “values.”

Promoted content that is deemed political will have a “Paid for by…” disclaimer, with a link to that ad’s entry in the database. From there, users can see the ad budget, the number of people who saw the ad, and basic demographics of those who saw it such as age, gender, and geography (by state). They can also see other ads from the advertiser.

How will Facebook know if my ad contains ‘political content?’

Facebook began enforcing the new policy on May 24, reviewing all new promoted content — specifically boosted posts and ads in the News Feed, Instagram feed, and Facebook Stories — for “political content,” rejecting those it considers political that do not carry the required disclosure.

To review content at scale, Facebook has created a “machine learning classifier” — in other words, a program will scan your content for identifiers of political- or issue-related material, and make a determination.

If the content is determined to be political, it will run the following checks:

  • is the person placing the ad authorized to buy ads with political content? (instructions below)
  • is the ad account linked to a Page with an admin authorized to place ads with political content? (instructions below)
  • when creating the ad or boosting the post, did the advertiser check the box stating “This ad includes political content”?
A slide from the webinar showing the political content checkbox.

If these three criteria are not met, the ad or boosted post will be rejected. Note that if the ad is not political content, and you have checked the checkbox indicating that it is, your ad will be rejected and you will need to resubmit the ad without the checked box.

Users can appeal Facebook’s determination, but should be aware that if a piece of content relates even loosely to the democratic process or the stated issues of “public importance,” Facebook will likely reaffirm its decision.

I’m a publisher. What do I need to know?

  • Get authorized. Here’s the how-to. Personal authorization does not expire, and is tied only to your personal account — not the ad account or the Page. Your entire team doesn’t need authorization, but only those who are starting, stopping, or modifying ads with political content.
  • Bookmark the list of issues of public importance. Working with third-party groups, Facebook identified 20 issues. They’re very broad, and the list is expected to change over time.
  • There’s a June 6 deadline. This is the day Facebook will unleash its machine learning classifier on all ads running on its platform, not just newly created ones. If you have ads that began running before May 24 and could be considered political content, those campaigns will be terminated unless you modify them and add the required disclaimer.
  • This only affects promoted content on the platform. It will not affect news stories or other content that you are posting to your timeline, regardless of subject matter.
  • Keywords are important. The machine learning classifier is scanning the text, images, and videos for terms related to politicians, legislative terms, parties, PACs, and issues of national importance. And these terms are broad, encompassing not just prominent political figures or issues, but titles and terms that are loosely related to them. A local news site writing “Legislative season reopens, representatives consider bills,” and does not mention the names of representatives or the nature of the bills, could possibly trigger the classifier.
  • It doesn’t matter if you do political news — or any news. The classifier is looking at the terms used in ad copy and the landing page, regardless of who the advertiser is. An ad for Arrested Development with the word “Trump” would likely trigger the political news classifier, as could a lifestyle publication running an ad about “health” — one of the 20 issues of public importance.
  • The classifier scans the linked page, too. It’s not just the ad content that Facebook is looking at; they’re reviewing any page you link to for political content. Facebook tries to look at only the main article content on that page, not linked content, sidebar content, or content in modules or advertisements. But the machine learning classifier is a work in progress; there’s an appeal process if you’re rejected.
  • The name of the ad account is who paid for the ad. The disclosure label will read “Paid for by [ad account],” and the disclosure statement is specific to the ad account. The ad account should make clear the source of the funding. If you’re promoting your own content, it’s “Paid for by [Publication name]”; if you work as an agency and run ads for clients, each client will need its own ad account and unique disclosure statement.
  • Facebook may treat news content differently in the future. Facebook representatives said they’re exploring ways to treat news content differently within the archive of ads with political content, and possibly within the flow itself. They have not made any decisions on this, however.
  • You cannot edit ads in the archive. Not for a typo, or an error, or any other reason. This implies that each revision of an ad will be stored in the archive, though Facebook didn’t say that specifically.
  • You need to use your personal address. The address for verification must be a residential address, not a business address. This is to confirm that people running ads are who say they are, and are based in the US. All identification data is used only to verify your identity and Facebook says it is discarded after that process is complete, in about 4–5 weeks.
  • This is going to be rolled out globally. If you’re outside of the U.S., be aware that Facebook is planning a global rollout of these rules.
  • Learn more. There’s a Blueprint course.

Was this helpful? Follow me on Twitter for updates and other audience development related tips and tricks.