Tuesday, June 3, 2014

Interest Wars: Facebook vs. Twitter

It'll Make Sense By The End Of The Post

Yesterday The Omnivore kicked off an experiment in social media advertising by analyzing both Facebook’s and Twitter’s information models around ads and then placing an ad on each of them. The learning: The Omnivore is a sucky ad writer. Here’s the breakdown to-date.

Facebook Twitter
Impressions 2592 18,321
Click-Through 35 4 (followers)
Cost $4.23 USD $6.12 USD

There was room for improvement. Let’s recap how each service fundamentally sees its users:
  • Facebook sees you as a person: you have behavior. You have complex demographics. You, then, have interests.
  • Twitter sees you as, not to be cynical, a collection almost purely of interests (you do have gender, a location, and a language—just like Facebook—but these are about the bare minimum statistics necessary to pitch an ad to you).
When The Omnivore looked at that, he realized he was making a fundamental mistake: the ad he’d used in both cases was pretty much separate from the targeting he’d done. Why?

The Omnivore’s ad was a specific post analyzing why Elliot Rodger used pistols instead of an AR-15 rifle: it got good feedback and a good number of hits and shares among The Omnivore’s native readers (if you are not liking and sharing The Omnivore, you are slacking!). So perhaps it was a good “ambassador” to The Internet. The problem was that it wasn’t—not for The Political Omnivore blog as a product, anyway.

That article might be of interest to people who were looking for psychological analysis of the shooter or his manifesto—but the top-line of the post read all about guns. Secondly, The Omnivore’s targets were political people—not people with an interest in the shooting, guns in general, or the psychology of shooters. This means that the specific product was a miss-match to the interests. Now, in theory, if someone read the article and then clicked around on the site they’d maybe go “Okay, this is interesting to me—I’m political … this site says it is mostly about politics … in a kind of RINO-ish fashion … I’ll come back).

In a world where nobody even finishes listening to a single song—much less reads an entire article through, The Omnivore’s fantasy of someone looking at the rest of the site ain’t going to happen.

Phase 2: Targeting Interests Instead of People

This brings us to the second phase of the exercise: targeting interests. Both Facebook and Twitter allow it—both with substantially different ecosystems of interest. NOTE: we are going beyond the actual taxonomy of interests both provide and into the world of ‘suggestions’ and free-form key-words. To do this, we need something to 'sell'--an "intellectual property" of some sort. Turns out? The Omnivore has one!

The Property: Illuminoimia


The Omnivore is writing and releasing a web-serial fiction story on the blog where every right-wing conspiracy is true. The name of the story is Illuminoimia and it is done in homage to The Illuminatus! Trilogy which postulated that every conspiracy theory of the 70’s was true. The world of Illuminoimia is the world of right-wing paranoia where governments are controlled by a shadowy super-conspiracy, Obama was born in Kenya, and Benghazi was about a US Government attempt to smuggle a nuclear weapon into Israel.

So what do we know about people who might be interested in a long web-story about the Illuminati? Three groups:
  • People, like The Omnivore who have an interest in conspiracy theory as an intellectual curiosity.
  • People who, unlike The Omnivore, think this stuff is maybe real.
  • People who want a fiction that vindicates the far right (Obama is a puppet, Hillary is running the government, and 2nd Amendment rights are all that stand between America and subjugation by the world-government).
How would The Omnivore find them?

It wasn’t in the lists of topics: there are no selections for any of that. Conspiracy Theory as a hobby doesn’t exist for either Facebook or Twitter as a specific selection.

Facebook Interest Targeting

On the other hand, you start punching ‘Conspiracy’ into the Interest search bar and you find all kinds of things. Here’s every Illuminati / Conspiracy related key-word The Omnivore could find for the Illuminati pitch:
These are probably both groups (actual groups you can like / join on Facebook) and more nebulous interests that Facebook has mined from keywords in your posts, web-searches, things your friends like, and so on.

Here’s the problem: it’s hard to know what, exactly, these mean. Illuminati has a lot of cachet in the rap world where it is (a) taken semi-seriously but (b) does not have the same meanings that it did back in the 70’s (although the idea that it’s primarily a Jewish conspiracy may actually be constant among conspiracy peddlers …). We also don’t know if these people Liked the pages that track this? Just visited them? Searched on them? These could all mean very different patterns of engagement.

The final ad looked like this:


Twitter Interest Targeting

There wasn’t a Conspiracy Theory category on Twitter either—and no random key-word search functionality either. So what did The Omnivore do? The Omnivore knows who conspiracy purveyors on the web are. He went to their twitter feeds! It’s easy enough to know where to start. Guys like Alex Jones and Glenn Beck are going to be attractive to the conspiracy hungry. Infowars has several twitter accounts … after that, we’re off to the races: Twitter can show you several pages (see below) of people who might fit your @username profile.
60’s Spider-Man?? Really?
Suggestions2This is where we finally wound up (Yes … Ted Cruz)

Now, The Omnivore doesn’t know all those accounts—Stop The Illuminati turns out to be a humor account around illuminati symbolism—but that’s probably still within Illuminoimia’s target market.
So here we go. This was the Twitter Final Ad:


The Results

Which performs better? Let’s take a look:

Facebook Twitter
Impressions 243 7937
Click-Through 5 Website Clicks, 7 “other clicks”
124 Clicks
1 Re-tweet
4 Follows
Cost $0.82 USD $1.40 USD

FacebookResults TwitterResults


In terms of driving followers, Twitter is far better than Facebook. If what you care about is ‘repeat customers’ (Followers, Likes) then the answer is clear: Facebook returned zero and only 5 people to the site. Twitter, for about 2x the price, got 4 follows, 1 re-tweet (which would be a Share on FB) and 124 clicks which might bring you to the site (they track any click on any part of the ad—it’s not clear to The Omnivore what, if anything, those do for you).

Why? Well, let’s look at some of those sign-ups. We kind of have to guess since it doesn’t tell you exactly who signed up and why—but The Omnivore thinks he can make some guesses (here are two):
Follow2 Follow1

If you look at the descriptions these are not generally The Omnivore’s target market—but they are Illuminomia’s target market. Compared to zero Facebook Likers. In other words, Twitter has done a substantially better job of targeting potential “buyers” than Facebook did.

The Omnivore thinks it is because the semantics of Liking a group-page are significantly different than Following a Twitter user. The differences are key and important:
  1. Liking may show support or interest Following shows pretty much only interest in what the party has to say. The Omnivore might Like his daughter’s high school page if she asks him: it doesn’t mean he care a lot about what they post. While there is probably some reciprocal Following, it is probably much more minimal when dealing with big names such as @usernames returns (Glenn Beck won’t Follow you just because you Follow him).
  2. Facebook content is often rich. This is nice—but, as you saw if you clicked the above links, people usually don’t read the full content. They may even comment without reading the article ( :: GASP :: ). Twitter tweets are 140 characters: You read the entire tweet in one look. This is a powerful difference in terms of “being heard.” I can post articles to Facebook all day and I may get Likes if people like the headline. When I tweet something, if it’s rendered on someone’s page and the look at it? They read the whole thing (Yes, The Omnivore is aware of multi-part tweets …)
  3. Facebook Friending in the first place is highly promiscuous. Twitter Following is less so. Friending is linguistically symmetric (you are friends with people who are friends of you). Following is linguistically uni-directional (if you follow someone in common language they are therefore not following you).
The results of these are that tracking interest by Twitter Follows is a very different beast than tracking interests by Facebook’s formula that is based on page-likes (and other data). In fact, it’s on the order of 100’s of %s better. It’s less user friendly (you have to find @usernames that apply to your product as opposed to key words)—but it’s far more meaningful.


  1. TL;DR:-)

    Where's my monthly check? Tell that to the "Worldwide Conspiracy"®!

    -- Ω

    1. Probably in a large, faceless corporate office's mailbox if I remember correctly ...

      -The Omnivore

  2. Anecdotally...I came here because you talked about Illuminoimia in a forum thread and just sort of stuck around.

    Talking yourself up online is a pretty good advertising tactic, I think.

    1. Talking oneself up *can* be good. On the other hand ... one's less likely to take it personally if your talking goes unrewarded. The Omnivore was originally going to compare: (a) Facebook (b) Twitter and (c) Begging for Likes, Forwards, Favorites, Re-tweets.

      He decided the results might be ... *embarrassing*.

      -The Omnivore