As a social media marketer, one of your primary goals is to distribute content across social channels to drive traffic to your site. You post content on various networks, try to build engagement and hope that it translates into page views. The problem is, the degree to which social engagement translates into page views has never been quantified.
After analyzing thousands of articles posted on Facebook, Twitter and StumbleUpon, we discovered the following:
- Overall, a 10% increase in social actions (i.e. shares, tweets, etc.) led to a 9.8-14.4% increase in social referrals. 
- Content posted on Facebook almost exactly mirrored this relationship since Facebook makes up such a large percentage of total social sharing.
- Interestingly, content posted on Twitter saw an even bigger lift – a 10% increase in social actions led to a 10.8-15.1% increase in social referrals.
What exactly does this mean for you? Optimizing your social media posts for higher engagement will drive more traffic to your site. You can follow a few basic tips to generate more engagement:
- Use high quality images and videos (especially with Facebook’s autoplay feature), as people are drawn to visual content.
- Insert hashtags whenever appropriate to show up more frequently in search.
- Keep posts brief for readability and to make it easy for users to share on networks with character limitations, like Twitter.
- Include a call to action (CTA) asking users to share your post.
For a more in-depth look into driving engagement, check out this excellent post by social media rockstar Jeff Bullas. You can also read this post for more tips on how to use data to inform your content strategy.
For those interested, here’s an in depth view of how we reached the above conclusions:
First, we looked at the ratio of social referrals to social actions. Unexpectedly, we found that the distribution of this ratio was not normal – the standard deviation was massive. This indicated a skewed distribution, and the histogram showed something much closer to a gamma distribution, primarily due to some extreme outliers.
To account for this skew, we used logarithmic transformation to help moderate the extreme cases. We also removed content with less than 10 social actions and social referrals to ensure there was enough activity to draw an accurate conclusion. Finally, we assumed that the extreme 0.1% of data points were outliers and removed them.
After these transformations, we saw a perfectly bell-shaped normal distribution with a standard deviation two orders of magnitude lower than the original histogram. Since the processed conversion rates were normally distributed, we could extract the results at a 95% confidence interval.
 It’s important to note these results are dependent on factors like the size of your social media audience and the type of content you produce.
About Adam Orshan
Adam is a Content Analyst at SimpleReach and tends to make somewhat obscure references.View all posts by Adam Orshan »