Want to get more shares with a headline? Optimize for this emotion.

Boost Linguistics released an emotional analysis function so I decided to test it out on 100 blog post headlines to see which emotion led to the most shares.

Want to skip my research methods and see the best emotion for a shareable headline? Skip to “The Results”

Why did I choose to only analyze headlines?

After some research, I came a across a very interesting stat that solidified my decision to analyze headlines.

“6 in 10 of you will share this link without reading it, a new, depressing study says” — Caitlin Dewey, Washington Post

I couldn’t believe it… 60% of people that read a headline will share it with their network before reading the article themselves. Damn. For a marketer this can mean one of three things. First, less people are going to actually read your content. Second, it is much easier to get shares and go viral because you only need 6–12 words to get a share. Third, the headline is the most important part, and you better put the most amount of effort into crafting it.

I’m a glass half full kind of guy, so let’s say its a trend that helps virality of content.

How did I measure shares of a post?

The blog I analyzed was Marketing Insider Group, they are one of the thought leaders in content marketing, and marketing practices in general.

Next, I needed a tool to analyze for the base human emotions, the Boost Editor would work perfect for this. The Boost Editor analyzes any piece of text for the base human emotions on a scale of 0-100% for each.

For example, if you type in “I love to go sailing in the summer.”, and analyze it, the results will be the following:

Emotional scores are each 0–100%

Then, I needed a tool to track shares of a URL on social media outlets. The numbers of shares on a post for the Marketing Insider Group is displayed on the article if it was shared more than 101 times. For all posts below the 100 share mark I used Sharedcount. Typically I use Buzzsumo, but I didn't need the in-depth data, just the shares.

Lastly, I needed an easy way to track the performance of headlines. Google sheets was perfect for this because anyone on my team could easily access the data.

This type of work is labor intensive, its nice to get some help 😊

There is so much more to a share than a headline…

I pondered the statement above before writing this article, but ended up deciding the 6/10 stat listed above was enough justification to only analyze the headlines of these posts.

The dataset

I analyzed the headlines of 100 blog posts from 3/1/2017 to 6/13/2017. I originally thought that the posts which have been up longer would have a higher number of shares, but it ended up not being enough to skew the data. There are a few constants in the MIG blog that I also chose not to analyze.

  1. Single book recommendations
  2. Infographics
  3. Links to podcasts

These three do not make up a significant number of the total blog posts. For every five or six long form blog posts, the MIG blog will have one of these three.

Hypotheses

Emotional marketing has become a trend as of late, and for good reason too. Recent studies by reputable outlets show emotional content can out perform regular content, leading to huge jumps in ROI, virality, and every other juicy KPI. One article I found by Noah Kagan shows that joy is among the top performing emotions for shares. Other outlets also backed up this data, so I added it to my list.

Graphic from Noah Kagan article

Awe, laughter, and amusement are not base emotions according to Ekman’s emotion spectrum, so I will consider Joy(14%) the best performing base emotion in the chart above.

  1. The higher the Joy rating, on average, the more shares.
  2. Joy is the best base emotion to optimize for to increase the probability of shares.

The Results

After analyzing 100 headlines, the range turned out to be significant. The lowest amount of shares was 24 and the highest 421.

#1 The higher the Joy rating, on average, the more shares.

In order to test this hypothesis, I organized the data in descending order from highest to lowest Joy rating. Then, I split the rating into 3 sub groups. First, less than 50% which was unlikely to elicit Joy. Second, between 51–75% which was likely to elicit Joy. Third, 76% and above which was highly likely to elicit Joy. Once the 3 categories were aligned, I averaged the number of shares received for each. Less than 50% received 127 shares. Between 51–75% received 135 shares. Above 76% received 167 shares.

From this analysis, I can conclude that the higher the Joy rating, the higher chance your headline will receive more shares on social media.

Joy rating in relation to number of average shares.

#2 Joy is the best base emotion to optimize for to increase the probability of shares.

To test this hypothesis, I collected the total number of shares when Anger was the highest rated emotion. Once I had the number of shares for each headline that was predominately rated for Anger, I took the average number of shares across every article. I then repeated this process for the remaining baseline emotions, Joy, Fear, Disgust, and Sadness.

Here is what I found, when Anger was the dominant emotion, the average number of shares was 83. When Anxiety(Fear) was the dominant emotion, the average number of shares was 65. When Joy was the dominant emotion, the average number of shares was 137. When Sadness was the dominant emotion, the average number of shares was 53. Disgust was never the dominant emotion in a headline and so the average was 0.

Number of shares when the emotion is dominant.

Lessons learned

There is much to be learned from these 100 headlines. The first conclusion I reached is, the rumors of emotional content out performing neutral content are absolutely true.

Second, the hypothesis: the higher the Joy rating, on average, the more shares. I found this statement to be true. On average, the higher the Joy rating of a headline, the more shares it will receive on social media outlets.

Third, the hypothesis: Joy is the best base emotion to optimize for to increase the probability of shares. This statement did turn out to be true. So much so that Joyful headlines were shared nearly 2x as much as the other emotions. Anger was the second best with an average of 83 shares. I believe that Anger has some potential to excite readers enough to cause a share from just the headline, but there isn’t enough information at this time to reach a conclusion.

For now, I’d say the safe bet when optimizing for shares is to make your headline as Joyful as possible.

Comments, questions, concerns? Tweet me @jefnwk

Show your support

Clapping shows how much you appreciated Jeff Nowak’s story.