The Half Lives of Social Content

Nick Law recently told me that while the web was about space, social media is about time. He was citing a piece in Wired called “The End of the Web” in which Dan Gelernter argued that streams of content have replaced the web’s “flatland”  where everything exists at a fixed position: a file path, or URL.

In the new timescape, content is always moving. We simply tune in to a social media stream and can watch in real-time as stories are generated. Publishing becomes so frictionless, we push back into the streams easily. Even without interacting, we find digital space consistently populated with new stories and data from connections known and new. As Gelernter imagines:

It’s a bit like moving from a desktop to a magic diary: Picture a diary whose pages turn automatically, tracking your life moment to moment … Until you touch it, and then, the page-turning stops. The diary becomes a sort of reference book: a complete and searchable guide to your life. Put it down, and the pages start turning again.

In this world of the magic diary, there are two key questions. First, how fast do the pages move? Second, what content breaks through the movement and makes us pay attention?

These are the current chief tactical questions of social media, where strategists and publishers look to understand stream algorithms in order to stand out in hurrying pace of data.

A Perspective on User Engagement 

In nearly every social media environment, user engagement drives the success of content. As a post receives likes/retweets/shares/reblogs/etc. it enters new streams where other users may too interact with the content amplifying its impact and visibility. On Facebook, user engagement proves particularly crucial, as it signals content quality to the News Feed algorithm and results in more or less reach.

With engagement driving so much, I recently decided to look at the comparative engagement time periods of social media content. Basically, how long does it take for a piece of content to generate most of its user engagements.

The current answers, culled from existing research and a small original study of Instagram, show the lives of social content to be very brief. Minutes, not hours are the primary units.

Tactically, the full lifespan of social media content is less interesting than its initial pickup.  Successful social posting is about something we could call “sharing velocity,” a measure of the speed in which early (read: impact) engagement happens. This is what promotes a story within the ‘fast flipping of pages’ and makes it stand out. It’s something we might call the “half-lives” of social media.

In science, a half-life is the period of time it takes for a quantity to fall to half of its value. If we imagine engagement growing towards a fixed, final engagement total, we can apply a half-life metric to see the moment it reaches half of engagements – a zenith in the user interaction curve after which it will likely long-tail. Half-lives of social media allow us to see how long something stays ‘fresh’ on a social media platform and when we might expect content’s impact to have passed.

Half-Lives of Social Media Content 

Ahead of Kristen Joy Watts “Sustainable Stories from Disposable Content” SXSW panel, I pulled together the existing data on the half-lives of social media content. The graphic design was done by the brilliant Elliott Burford.

HalfLivesOfSocialMediaContent

Sources

A lot of research had already been done on the half-life of engagement for Facebook and Twitter. In November, SocialBakers reported that a post will receive half of its lifetime engagement in the first 30 minutes its published.  This followed earlier research by Bit.ly that showed half of a Facebook post’s clicks coming in 3.2 hours. But a major change in the Facebook algorithm separates these two posts and likely has shortened the Facebook post engagement curve.

Bit.ly research also showed the average Twitter link reaching its half-life in 2.8 hours from their study in June 2012. But a November 2012 study found that half of all retweets came in just 18 minutes. From experiences, that seems most right as content moves hyper-fast in a Twitter stream.

The YouTube data, pulled from Bit.ly, is probably the diciest data. Because YouTube works well as an archive, a piece of content can always go viral well after being published. So while Bit.ly may be right about click-through YouTube links, I am not sure about all-time watches.

Original Research on Instagram

To date, there has been no public research in how long it takes an Instagram image to reach its engagement half-life. So along with Kristen Joy Watts, I conducted a small study via Statigram. We asked two dozen friends and colleagues to examine their “Photo Average Lifespan” under Statigram’s Statistics > Optimization menu. This lifespan measures photo comments over time, a good indicator of inter-user engagement even while likes over time would be much preferred.

With nearly 18 responses, the clustering was very consistent. Instagram users surveyed received nearly all comments within 5 hours. The range of most comments was from 3 – 6 hours overall. Averaging the lifespan numbers, we then divided by 2 to find the half-life. The average instagram image receives half of its lifetime comments in just 2.23 hours!

It’s Quickening 

The lifespan of social media content is shortening. Arranged chronologically, each article would show the period of prime user engagement shorter and shorter. There is also, of course, the dimensions of user influence and amount of followers/friends to consider. Klout did a very interesting study of Twitter content lifespan correlated to level of influence. Unsurprisingly, the influential user’s content lasted longer.

Viewed optimistically, the shortening half-lives of social content is not a demise of cultural value but a new opportunity. We increasingly realize that the most impactful social posts are timely. Or they come from a place of habit: a creative formulation where repetition brings emergent value. Much like Maddie on Things, we can learn to love the frequent return of light, original content in the rapidly flipping pages of our timely web.

This entry was posted in community management, Culture, data culture, Internet, new york, research, Social Media, Social Networks. Bookmark the permalink.

3 Responses to The Half Lives of Social Content

  1. Meaghan says:

    Hi Zack,

    This is a great post! I had a question about how you calc’d the half-life of the Instagram photos… you said you divided the average lifespan in half… if the growth of the Instagram comments is linear, then this makes sense (half the lifespan aligns with half the total number of comments); however, if the growth in comments is exponential instead, as implied by the idea of a half-life (or perhaps better yet, doubling rate) then I’m thinking this might underestimate it.

    For example: If you have 100 comments after 5 hours and you the half-life/doubling rate is 1 hour, then it will take 4 hours (not 2.5 hours, which you would get by dividing the total time by 2) to reach 50 comments, or half of the total.

    To figure out what the half-life/doubling rate, all you need is the number of comments at two different time points (two snapshots of the comments as they grow) and then you can figure out the rest of a formula like this one:
    http://en.wikipedia.org/wiki/Doubling_time

    Anyways, I found this post super interesting, and I think your concept of an “engagement half-life” metric could be very insightful.

    All the best :)

  2. Great Post Zach – I’m only just discovering it …

    I wonder if the longer term impact of Pocket and Buffer’s presentation of seemingly native controls onto pages for adding to your queue might lead to more usage or if these will stay in the power user focus for the foreseeable future. Filtering the signal from the noise is a great motivator, but only if you consume regularly. It’s not something my wife for example who reads twitter to supplement her weekly gossip flow will really care about since the content expires so quickly — which is key detail here. The average tweet or instagram post doesn’t necessarily warrant a save later.

  3. sy says:

    And: “In the western Deserts, tattered fragments of the Map are still to be found, Sheltering an occasional Beast or beggar…”

Leave a Reply

Your email address will not be published. Required fields are marked *

You may use these HTML tags and attributes: <a href="" title=""> <abbr title=""> <acronym title=""> <b> <blockquote cite=""> <cite> <code> <del datetime=""> <em> <i> <q cite=""> <strike> <strong>