I decided to begin my analysis by using a tool called Wordstat to help create a taxonomy of BuzzFeed content based on their titles, which would then give me the ability to understand the associated data of each article in a more manageable way. I believe the categories I arrived at are comprehensive and represent, in broad strokes, the multiplicity of ways BuzzFeed thinks about content creation.

To further elucidate the findings, I segmented these content categories into two major groups: a Topical Grouping and Qualitative Impact grouping.

• Topical Groupings – categories related to a person/place/thing/idea
  • Numbers
  • Food
  • Social Media
  • Years
  • Locations
  • Women and Women’s Issues
  • Politicians
  • Movie Fandoms
  • Consumerism and Products
  • Female Celebrities
  • Sex and Beauty
  • Music
  • Men and Men’s Issues
  • TV Fandoms
  • Animals/Pets
  • Title Includes Internet Slang/Jargon
  • Title Includes Swear Words
  • Title Includes Superlative Words
  • Photos/Photography
  • Brands
  • Hacks/Tips/Tricks
  • LGBT
  • Male Celebrities
  • Relationships
  • Superhero Fandoms
  • Memes
  • Cuteness
  • Race/Ethnicity
  • Sports

• Qualitative Impact – categories related to how the article intends to make you feel or respond.
  • Articles that convey the emotion of trust or authoritativeness
  • Articles that convey the emotion of surprise or anticipation
  • Articles that convey the emotion of anger
  • Articles that convey humor or are meant to be funny
  • Articles that convey the emotion of joy/happiness
  • Articles that create a sense of collective identity
  • Articles that convey the emotion of fear
  • Articles that convey the emotion of disgust
  • Articles that convey the emotion of sadness
  • Articles that relate to the aspirational self

These groupings will be analyzed independently in order to allow for better “apples-to-apples” comparisons between categories.

Before we delve into these groupings, let’s first take a look at the distribution of articles across all categories, in order to answer the question, “Which categories do BuzzFeed and their audience seem to favor most?” What follows are the percentage of articles from the corpus (more than 15,000 articles) that fall into each category. It’s important to note that many articles fit into multiple categories, sometimes even more than 5 categories.

Takeaways:

1. Article titles that include numbers—commonly known as “listicles”—are the most common category by far, with almost half of all BuzzFeed article titles incorporating a number in some form.
2. This visualization makes it clear that BuzzFeed’s highest priority has not historically been hard-hitting content; instead, they have found success with topics like food, animals, celebrity gossip, pop culture, and social media commentary.

Given the disproportionate favoring towards some of these categories, it’s clear BuzzFeed is very adept at leveraging what they know works. BuzzFeed publisher Dao Nguyen has spoken about how the media company has drawn on broad thematic goals for its content production, including the following groupings:

• Humor: “Makes me laugh.” There are so many ways to make somebody laugh. You can be laughing at someone, you could laugh at specific internet humor, you could be laughing at some good, clean, inoffensive dad jokes.
• Identity: “This is me.” Identity. People are increasingly using media to explain, “This is who I am.” They use relatable content to share specific aspects of themselves, such as their
  upbringing, culture, fandom, guilty pleasures, self-love, and sense of humor. This type of digital connection is one of the greatest gifts of the internet. It’s amazing when you find a piece of media that precisely describes your bond with someone or something.
• Practical/Useful: This group is characterized by helpful content that allows someone to complete or solve a task, such as DIY, shopping and reviews, learning and education, editorial/news, etc.
• Emotional: Feeling, feeling, feeling. These posts have the burden to create an emotional response, evoking feelings such as sadness, curiosity, happiness, anger, and despair. Often times, these articles are raising awareness of a problem facing humanity or sharing news that is meant to restore faith in humanity.

As mentioned, we attempted to go deeper here with our analysis, parsing the top 15K articles into more specific categories within the broader “buckets” listed above.  Using these categories, combined with information on Facebook’s sharing and linking data, some very fascinating trends emerge that have practical implications for all content creators. Some categories of content perform disproportionately well on these important metrics.

Next, we will be looking only at the articles that could be considered “top performers,” having earned more than 50,000 Facebook shares.

Let’s take a look at which categories are represented by these widely shared posts:

Takeaways:

1. Less than 3% of the top articles we explored earned more than 50,000 shares.

Articles with titles that include words referencing someone’s identity/collective identity and memes are disproportionately represented by a higher than average number of shares. More than 15% of the articles in either of these categories garnered more than 50,000 Facebook shares each. Meanwhile, content related to categories such as Movie Fandom, Race/Ethnicity, and Politicians struggled to compete.

People are more likely to share things that they think their friends will find interesting and relatable such as Identity and Internet culture, rather than polarizing content.

Kristin Tynski, Fractl

 

Shockingly, headlines mentioning Sex/Beauty and Female Celebrities had a smaller representation of posts earning more than 50k Facebook shares. This could because people are turning to other sources, like Instagram, Pinterest, or YouTube for their content in these areas.

Topical Categories Group

Now let’s isolate just the Topical Groupings:

Major Takeaway:

1. Excluding the “Qualitative Impact” grouping, we are left with a good impression of which topics seem to generate the highest levels of social sharing. In particular, somewhat generic aspects of “Internet Culture” seem to dominate, such as Memes and Internet Slang/Jargon. Pets/Animals content also fared well, with more than 12% of related posts garnering more than 50,000 shares.

Lastly, let’s look just at the categories relating to Qualitative Impact, specifically categories having to do with identity and specific emotion:

As I’ve investigated in past studies on the emotions of viral content, some emotions have been shown to elicit higher rates of sharing. In this case, the top emotion-based category was not titles that conveyed anticipation, or “Surprise” as we’ve seen with other investigations in past research (though this emotion did beat out Disgust and Anger). The top-performing emotional categories were article titles that conveyed sadness or humor.

However, these emotional categories were dwarfed by any titles that leveraged feelings around individual or collective identity. Perhaps this is a reflection of just how powerful identity has become in the current social conversation. It could be that sites like BuzzFeed are intentionally creating content and using titles meant to stoke feelings of in-group/out-group and identity stratification because they know how effectively this creates social sharing behavior.

Facebook shares are important and interesting when considering social dialogue trends. But for the SEO-driven content marketers out there who are thinking about optimization, how do links factor into this research? Which types of content drove the most number of new links for BuzzFeed?

Interestingly, we observed that the categories that performed well with Facebook shares did not always translate into high link drivers.

Takeaways

1. Even though headlines related to politics weren’t heavily shared on Facebook, other media outlets online were using BuzzFeed as a source and linking back to their content. In 2018 BuzzFeed made more efforts to be seen as a news source, and even had their work on Russia, and operatives with ties to Vladimir Putin taking part in targeted killings against Putin’s enemies, make it as a Pulitzer Prize finalist.
2. We see categories like Race, Male-related headlines, and surprising content earning links but not social shares. This could be because people are less likely to share polarizing topics on their Facebook feeds, but publishers are using posts like these to start conversations.
3. Location-based content performed relatively well in link driving, despite being the second-lowest shared category overall across all title groups. This could be due to regional media publishers picking up on BuzzFeed content for news stories.

BuzzSumo reports an “evergreen score.”

“The Evergreen Score is an internal ranking system developed by BuzzSumo. It measures the number of social engagements and backlinks an article receives 30 days after the article is published. The more engagements an article receives after the initial 30-day period, the higher the score.”

Using this score, along with the article categorizations, it’s possible to understand which categories drive the future value in terms of links and social shares. The higher the score, the better the “longevity” of the content in that category.

The big winners we see with higher than average overall evergreen scores range across a variety of verticals, such as Collective Identity, Race/Ethnicity, humorous memes, Animals/Pets, or self-help tricks and life hacks.

It’s also interesting to look at the distribution of the articles by type, meaning the intent and organization of the article itself. This could mean the questions it seeks to answer or the overall goals of the piece. These groupings were determined by BuzzSumo.

Categories can also be broken down into these “content-type” buckets. So, what can we learn from these views?

Lists are a huge priority for BuzzFeed. Roughly 40% of all the content they create exist in some form of list format. Lists are broken up by easy to follow sections and are becoming an increasingly popular way to present an idea.

Some categories are highly disproportionate by type, and the List articles are no exception. For instance, article titles with Numbers (of course), Humor/Humorous topics such as Memes and Swear Words, and helpful topics such as “Things I Can’t Live Without” rank as the categories
that use Lists the most.

Categories around Celebrities, Politicians, Fandoms, and Race/Ethnicity show up as categories that use “General Articles” the most, likely because these categories require more story development, researching, fact-checking, and writing to make their point of impact.