March 9, 2011 12:09 PM   Subscribe

The Like Log Study: [SLVimeo] What can we learn from Facebook reactions to online news? Sortable statistics from a study on Facebook "Likes" of major news sites and stories.
posted by Fizz (11 comments total) 5 users marked this as a favorite
So where do I click to Like that guy's name?

Kneejerk aside, the data is interesting to see (especially the "heat maps") but not terribly surprising. The topics seem to line up well with the activity I've seen here and on other sites.
posted by The Winsome Parker Lewis at 12:26 PM on March 9, 2011

Could we get the actual link put in the FPP, and remove the shortener? Here is the link.
posted by hippybear at 12:37 PM on March 9, 2011

From the Top 40 articles: Tired Gay succumbs to Dix in 200 meters.

Of course it's right near the top!
posted by ericb at 1:08 PM on March 9, 2011 [2 favorites]

Hey guys, thanks for sharing!

I am the author of the study, happy to take any questions. Just paid 5$ to to jump in :)
posted by yurylifshits at 1:09 PM on March 9, 2011 [2 favorites]

2ericb So it was Metafilter's job to promote Tired gay story?
You guys can turn an article into the real hit :)
posted by yurylifshits at 1:11 PM on March 9, 2011

"Sadly, web articles are basically lost 24 hours after publication. Our measurements show that fewer than 20% of likes come after the first day."

I suspect that the majority of MetaFilter FPPs follow a similar trend line.
posted by ericb at 1:14 PM on March 9, 2011

What's interesting about the "lost after 24 hours" thing is, when I look at the Google News page, which I use as a kind of headline aggregator for myself, I see articles, not stories, but the exact same URL, remaining on that page sometimes for 2 or 3 days. Especially in the Health section, and in the "Popular" feed, which is somehow automagically created out of some kind of data about the article being shared.

Does this mean that Google's screwing up? Or that there are only a few articles which really cross the threshhold into being "Popular"?

I know this study didn't have much to do with Google News or anything, but it's sort of an interesting counter-example, at least what I've seen there. (Anecdote not being data, I know.)
posted by hippybear at 1:20 PM on March 9, 2011


By 'lost' I mean 'lost by human readers'. And I only claim that reactions have stopped. It can still be in search results, there are still new visits, but no new tweets and no new likes.
posted by yurylifshits at 1:26 PM on March 9, 2011

Hey Yury, thanks for stopping by! How did you determine that there are 10 likes per 1000 pageviews? Also, do you know the ratio for pageviews per retweet?
posted by joedan at 3:07 PM on March 9, 2011


Several websites such as Forbes Blogs, Gawker Network (Gizmodo, Lifehacker, Gawker), Business Insider and display individual pageview counts for every article. Using Facebook API and TweetMeme API i retrieved the like/tweet counts for all their articles. Then I took an average value of (like count/pageview count). I computed separate averages among (1) all articles (2) all articles excluding top 10.

Here are the results:

In fact, this chart is on the original study page:
posted by yurylifshits at 3:13 PM on March 9, 2011

When I hear about "liking" a news story, I am reminded of a conversation with an employee of Preview House, which set up a dedicated theater in Hollywood in the '60s for pre-testing TV shows - and TV commercials. He said they set up dials in the armrests of the first two rows so viewers could give instant reactions to whether they were interested or entertained by what was on the screen. The test-givers totally emphasized that it wasn't about 'liking' anything, it was about how it held your attention. Still, people being tested always turned the knobs down when the villain appeared or something bad happened to the hero. He was positive that testing was responsible for the proliferation of long 'training montages' in the '70s/'80s... because the one in "Rocky" rated through the roof. I remain suspicious that a lot of world news is de-emphasized because the foreign despots test badly.
posted by oneswellfoop at 4:36 PM on March 9, 2011

« Older Iceland   |   What metal(s) do you want in the new coins? Newer »

This thread has been archived and is closed to new comments