The Science of Social Media: How Anti-Engagement Works

Yesterday I attended Dan Zarella’s “Science of Social Media” Webinar, where he demystified plenty of assumptions in social media marketing effectiveness.

Old laboratoryHere are the 5 facts that surprised me the most. But before I continue, a couple of caveats:

a) Dan claims that his assumptions are based on correlation and not causation.

b) These discoveries not necessarily go against best practice , they simply aim to prove what works and what doesn’t in terms of reach.

5 interesting claims:

  1. Engaging in the conversation doesn’t result in more reach, sharing interesting content does (links). Dan run an experiment confirming that accounts with under 1000 followers are more conversational while accounts with more than 1000 followers tend to share more links. He also got admin access to important Facebook accounts, confirming that loads of commenting didn’t result in more views but sharing loads of interesting links did. I assume this has to do with people re-sharing interesting links and this gaining more visibility in Facebook compared to the visibility that one’s own commenting gets towards your network.
  2. More blog commenting does not translate into more views or visits.  Again, another experiment analysing popular blogs from an admin perspective and finding no correlation. This one is a dubious one in my opinion, especially in relation to the commenting system the blog uses. For instance, if using Disqus or Intense-debate, which allow users to identify themselves with their Facebook account, their comments gain visibility in their Facebook stream too. Commenting system aside, I’m also not sure in statistics terms what level of sampling would be needed and how many assumptions should be made to reach something like this. It’s good to consider it anyway.
  3. Taking advantage of “information voids” is effective: Dan suggests that researching unanswered questions and producing content to tackle them work. It sounds obvious, but how many of us do it? Do you have a Twitter column open with your typical industry keyword and a question mark so you can serve people?
  4. Friday’s and weekends can be good days to publish:  This is another logical one but not everyone takes advantage. It’s simple, during weekends people receive less email and streams are less cluttered, therefore you can typically have more visibility (please don’t discontinue your work-day efforts though!)
  5. Social proof and novelty are different (my favourite one): Dan ran an experiment with 2 given blog posts, altering the “tweet button” and running an A/B test. In one version the tweet button was fixed to show zero tweets. In the other version it showed 776. Dan (and I) would have assumed that the one showing more social proof would be shared the most but it is actually the contrary. What one needs to understand is that social proof is good for mitigating risk (you choose the restaurant full of people over the empty one to reduce the chances of an awful dinner). Once you understand this, you need to combine it with “why people share things”, which is partly because they want to be seen as a reference/ authority. So using a tweet button has probably not much to do with risk and more with our egos. Bottom line: When it comes to sharing, nobody wants to be 777th person breaking the news. You could argue that this is valid for breaking-news-content and not for more perennial content and you would probably be right. My takeaway is that we need to experiment more with sneak-previews to a privileged segment, if we want to maximise the chances of our content being shared. We also need to offer social proof in more clearer risk scenarios (e.g. email subscribing), where phrases like “1000 people can’t be wrong” may mean a lot to our visitors. 

This is just my selection of interesting facts. You can grab Dan’s latest book here, which is free for Kindle users for a limited time. I’m nearly half way through it and I’m liking it, even if I don’t agree with some statements. But that’s the way it should be right? Regardless, I think the book makes a tremendous effort in moving away from a lot of rhetorical industry talk, which is something anyone with a bit of “social media saturation” requires :)

Have a look at the full presentation below and please share your thoughts.