Use Social Media to Iterate on Content Quality

Posted on September 11, 2012

What is social media good for? Absolutely nothing? When measured directly social media tends to perform poorly when compared to other channels but to some this is an essential channel with obvious benefits.

People who evaluate social media on a direct response basis have a model in their head like this:

digraph G {

But I think that if you took a company that was absolutely brilliant at engaging their audience through social channels and then banned them from Facebook, Twitter etc. they would still do well. These companies have the rare gift of being able to create content that engages their audience and that is valuable even without social media.

So now we have a model like this:

digraph G {
  social->revenue [style = dotted label="weak link    "]

“But wait!” you say, “is there really a such a strong link between good content that does well on social media and good content that generates revenue? Last time I looked squeeze pages were not the great content that everyone talks about”

Mostly the people talking about great content in the little filter bubble I live in are SEOs and they like great content because people link to it.

digraph G {
  social->revenue [style = dotted]
  content->revenue [style = dotted]

Now social media looks pretty cut off from the actual money side of things; this is not the point I’m trying to make. To get to the conclusion I need to make one big assumption:

People who link to stuff and people who share things on social networks like the same type of content

To me this seems like a reasonable assumption, but I’m still looking for data to back it up (or contradict it).

Please notice that I’m not saying that social media shares caused links or that links cause social media shares. I’m saying that the rate at which people share and the rate at which people link are correlated because of a common cause; the quality of the content.

Linking is a comparatively rare thing and you can expect link rates of below 0.01%. It is touch to optimise for an event that is so rare: You end up never getting meaningful results because the sample sizes are so small.

When this happens with a conversion event rather than a link event (for example if you sell something really expensive) people optimise for micro-conversions; proxy events which occur more frequently but which have an affinity for the main conversion event.

I propose that social shares are a useful proxy metric or micro-conversion for the main event; getting a link.

If this view is accurate it has the following benefits:

Being able to rapidly iterate on ideas is key for online businesses. By using social shares as a proxy metric the same processes that produce winning PPC campaigns and killer landing pages can be used to create content that engages the audience you want to engage with.