This is part two of a two-part series evaluating the practice of soliciting shares in your marketing content.

In part one we presented data from a survey we ran that looked at the pervasiveness of content creators soliciting shares, and what consumers thought about those solicitations. This week, we’re looking beyond the anecdotal consumer opinions, to more concrete data where we analyze hundreds of actual blog posts on Medium.

Two questions we answer in the second part of this research, that we expect will be helpful to you as a marketer, include:

  • How prevalent is the practice of authors soliciting their readers to share the content?
  • When authors do request readers to share their content, does it lead to a higher likelihood of that content being shared?

Background

Before we answer these questions, though, we need to give some quick and light background into this topic. If you’d like to see all the nitty-gritty details of how we approached this research topic, please see the research process and assumptions section at the end of this article.

Content marketers work so hard at creating content that we can assume it is one of their top goals to have as many people read that content as possible. In comes social sharing… If an author thinks that they can get more eyes on their content, if they ask their readers to share it, then they might solicit a share request, right!?

In last week’s survey response data about whether you should solicit social shares in your content, we looked at several means whereby an author might request social shares, including Facebook “thumbs up” solicitations, blog embedded “Click to Tweet” solicitations, and “clap” requests on Medium.

Here is a quick recap of several interesting finds in last week’s survey data:

  • 82% of respondents indicated they see a “share this content on social media” type solicitation at least once a week
  • The most commonly seen solicitation is a “thumbs up” request on Facebook
  • Only 12% of respondents indicated they had shared content via one of the solicitation techniques we were evaluating
  • The social share solicitation that most annoyed our respondents was when marketers challenged their readers with a comment such as “I dare you not to share this!” with respondents pegging this technique somewhere between very unfavorable and unfavorable.

For this week, why did we focus our analysis on Medium posts?

Analyzing social share solicitations on blog posts on Medium.com
To ensure data was as bias-free as possible, Ben Beck and a partner at Penn State evaluated 736 different publications. From those publications, they randomly selected 6,527 posts to evaluate at a high level. For deeper data on social share solicitation, they then randomly selected 500 posts to evaluate at a very low level.

While we believe that the findings in this research are generalizable to multiple content distribution channels, such as Facebook, Twitter and private blogs, we’ve focused this second part of our research on Medium.com hosted blogs. If you’re interested in seeing why we chose Medium for this endeavor, please be sure to read the thrilling (yay!) research process and assumptions section at the end of this post.

As a reminder from last week, below is a common style of a social share solicitation you might see on Medium, asking the reader to “clap” to indicate their enjoyment of the content – a measure that the Medium has explained will help get that blog post more attention in their network of content.

Example of a "clap" social share solicitation on Medium.com

The data helps answer: Should you solicit shares in your marketing content?

In a non-data-driven analysis, simply looking through our data without having run any analysis, it looked like there was no direct correlation between soliciting shares and receiving them. For example, this post about key elements in building a corporate startup solicited a “round of claps to keep [their] stories coming!” but had received no claps at the time of this article being written.

Give a round of claps to keep these stories coming on Medium!

And yet, this post about JavaScript programming has garnered 930+ claps with their fun social share solicitation, “If you found this article classy and/or you’d like to show the love for cats, don’t hesitate to hit some claps.”

If you found this article classy and/or you’d like to show the love for cats, don’t hesitate to hit some claps.

What the data says

If I wasn’t the geeky marketing technology lover that I am, I might have given up at this point, assuming that there was no correlation between soliciting shares and having actually received them. As there are many different components of a blog post (its length, whether the author used pictures, etc.) beyond just a social share solicitation, I wanted to evaluate the data in a more systematic way.

 

Percentage of posts with social share solicitations and those requesting claps on Medium.com

The two questions I wanted to answer with this analysis are below, with my best answers.

 

1. How prevalent is the practice of authors soliciting their readers to share the content?

I was quite surprised to see that not many authors were soliciting social shares of any kind. In fact, only 13.3% of the posts we evaluated had any sort of social share request. Of those that did make some sort of social sharing solicitation, less than 1/3 were requesting a clap. As a whole, looking at all the posts we evaluated, only 4% requested claps.

Realizing that so few authors were requesting claps, in particular, helped me see the need to expand my research parameters further for the next question.

2. When authors do request readers to share their content, does it lead to a higher likelihood of that content being shared?

Instead of looking at just share requests, which had a very small likelihood of occurring (less than 15% of posts request a social share), I needed to expand the research framework to also look at other potential variables that may influence the number of claps a post was getting. In the expanded model I also evaluated the length of the post and an author’s use of pictures, videos, headlines to break up the content.

I first ran a regression analysis looking at just one factor (whether they requested claps) and one response (whether they received more claps). This first test came back with no statistically significant results: Asking for claps on Medium does not increase the likelihood that you will receive claps.

Next, I ran a regression including two factors (whether they requested claps and whether they requested any share) and one response (whether they received more claps). This test also came back with no statistical significance but did present an obvious finding: that asking for claps had a higher correlation to receiving claps than asking for a social share of another type.

Lastly, I ran a regression looking at all the variables together, in hopes that some of the inner variable interactions may explain enough of the actual result (getting claps) to have statistical significance.

In evaluating all of these variables there was a little higher correlation assigned to certain variables, but only one was somewhat significant: Utilizing a greater number of internal headlines to break up your copy increases the likelihood that you will receive more claps. The level of statistical significance here is just 80%, meaning that we are 80% confident that the number of internal headlines you use is correlated to the receiving more claps for your blog post on Medium.

The takeaway?

From the data, there does not appear to be any statistically significant correlation between requesting claps at the bottom of your Medium blog post and receiving them. However, the data did show a different finding, that using more internal headlines to break up your content does have some correlation to garnering more claps on Medium.

It must be remembered, though, that correlation does not infer causation. Perhaps the reason why these Medium posts with many internal headlines were getting more claps was due to the overall adeptness of the authors to both write better content and organize it better.

Research process and assumptions

The reason for using Medium hosted blog posts for this more data-driven analysis is that Medium provides a unique, central repository of a vast array of topics. This provides an advantage to us as we were able to capture lists of the most popular Medium publications and then randomly select posts from a variety of publications, industries and reader targets.

In short, to ensure our data is as bias-free as possible, we’ve evaluated 736 different publications. From those publications, we randomly selected 6,527 posts to evaluate at a high level. For deeper data on social share solicitation, we then randomly selected 500 posts that we evaluated at a very low level. We manually viewed each one of those 500 posts and eliminated 11 blog posts that were not true blog posts and would be considered spam or irrelevant to this study.

Of the remaining 489 posts, we evaluated the following data points: length of the post (Medium reported “minutes to read” and the actual word count), the number of graphics/pictures used, the number of videos used, the number of internal headlines the author used, how many comments the post had, which tags were used to describe the nature of the content, whether the author solicited a share request in the content, and finally, the variable of most importance for this study, how many “claps” the post had.