Evergreen content is the Holy Grail for some online content creators and publishers. For those who do not know what “Evergreen Content” is, it is basically the content that blossoms (drive referrals) yearlong, and thus the name “Evergreen”. The goal of this blossoming content is to get lots of engagement in the form of activity on social networks (likes, shares, pins, etc) and page views. The Internet is not short on advice for producing evergreen content either. But at SimpleReach, we feel that given the wealth of data that we collect, we are in a special position to throw in our $0.02 on strategies to producing evergreen content.
While evergreen content can be one of the ultimate goals for content creators, it is extremely rare to come across a piece of content that is highly engaging year round. Since most successful content usually has on and off periods of high activity, we decided to name these periods of high activity a Life. Throughout this post, we will use SimpleReach data to provide interesting insights about the behavior of socially engaging content and how it affects their lives.
We began this process by taking roughly half a million pieces of content from 13 of the top publishers within the SimpleReach network. We were extremely careful in choosing the content to use for this analysis. In order to make the results reliable, we needed to ensure there was no selection bias. To achieve that, we made sure to include all different types of content ranging from articles to videos to images. We didn’t stop there either. We also were sure to include a wide range of topics including business, news, lifestyle, entertainment, and sports.
Leading a Life vs. Just Being Alive
In a previous post we discovered that most of the social traffic for publishers is driven by a small percentage of their overall content (“The top 1% of articles, ranked by social referrals, drove 67% of total social referrals across the SimpleReach network.” ). We identify those pieces of content that contribute the most social activity as Leading a Life. Everything else, or the content with minimal social activity, is identified as just Just Alive. To define a Life, we say that it is the period of time where the social activity for a given piece of content is above average for the publisher that created the content. Since Leading a Life is not always completely black and white, ups and downs are to be expected. Therefore, a content item that Leads a Life might actually have several lives over time. Here are some examples of Leading a Life and being Just Alive in Figure 1 and Figure 2, respectively. These figures are interactive, you can select between items using drop-down-menu as well as zooming in using the brush in the lower part of each figure, so have fun exploring.
Figure 1: Examples of Just Alive Content (No Life)
Taking a closer look at Figure 1, none of the social activity (i.e. social actions) exceeds the publisher’s normal social action traffic (red curve). Even though there is some activity, none of it is significant enough to be anything other than Just Alive. Figure 2, on the other hand, shows content with one life (shaded areas) that lasts for a longer period of time; more than two weeks in some cases. Specifically what we are seeing in Figure 2 are some examples of trending articles that drive a significant amount of social actions. It is also interesting to see how content transitions into being trending or having a Life. You can play around the figures to get a better idea of how that works.
You may notice that the lives of some of the content items in Figure 2 include areas where the social activity falls beneath the publisher’s normal social activity and yet we include them as part of the life. One of our interesting observations is that it is unusual for a piece of content to receive social activity every hour. And since there seems to always be periods of decreasing activity due to the natural cycle of human activity, we had to use a filtering rule to compensate. The filtering rule says that activity can still be considered part of a life as long as it lasts less than 5 days (we will cover where the 5 days comes from later in the post).
Figure 2: Examples of Leading a Life Content Items (One Life)
Now that we know the meaning of life (see what we did there), it is time to see how publishers compare. In Figure. 3 we provide statistics of the Leading a Life (i.e. at least one life) versus Just Alive (i.e. no lives) for our 13 publishers. We can see that the percentage of Leading a Life (i.e. trending) content items averages to roughly 10% across the half a million content items we used.
Figure 3: Statistics of Leading a Life (i.e. with at least one life) versus Just Alive (i.e. with no lives) contents for SimpleReach Top Publishers
As we mentioned earlier, many content creators strive to produce evergreen content. Sometimes the next best scenario, and generally the more realistic one, is for content to resurrect after its first life. This means that content will find a Second Life or maybe even multiple subsequent lives. This gives publishers a way to maximize their returns on content as long as they know when a new life for a piece of content begins.
Moving on to Figure 4, we have several examples of content with multiple second or more lives. The examples in the Figure 4 cover a variety of cases which include first and second lives, the total number of lives and their duration, the periods of time between lives, and early start versus delayed lives (see Leading a Life 3 in Figure 4).
Figure 4: Examples of content items with multiple Second Lives
Earlier in the post, we discussed the idea that a life cannot be one continuous period of time. This led us to create a filter that said if two lives are less than 5 days apart, then they are actually part of the same life. We figured this out by testing the sensitivity of the number of detected lives to the number of days between any two lives. We actually have the results of those comparisons in Figure 5. You can see, as we did, that the rate of change begins to stabilize around the 5 day mark. And since we know from previous observations that social activity has a tendency to decrease on the weekends, we felt that there was enough information here to support our theory that 5 days is a good threshold. Using a threshold of 5 days gives us the ability to ignore a weekend (or even a long weekend) with ample opportunity for the numbers to stabilize.
Figure 5: Rate of change of number of lives with different thresholds for the number of days in between lives.
Now that we’ve spent all this time setting up what a life is, looks like, and how long it lasts, let’s take a look at how many of these we actually see. Figure 6 provides a breakdown for each publisher into content with one, two, or three or more lives. We found that, on average about 97% of all the content considered to be Leading a Life lead just one life. Content that leads 2 lives accounts for about 2% of content that Leads a Life. And lastly, content that leads three or more lives accounts for only about 0.5% of all content included in our experiment. This is a testament to how difficult it can be to actually create evergreen content. On the flip side, it is also exposes how valuable it can be to know when a piece of content is getting a new life and take appropriate action to maximize its value.
Figure 6: Statistics of Leading a Life in terms of those with 1 life, 2 lives, and more than 3 lives.
One of the most interesting findings we have about content with multiple lives is that the duration of the second life is greater than that of the first life.
Figure 7: First and Second Lives Time Durations for Different Publishers
Second Life Drivers
At this point, the question on your mind is probably, “how do the various social networks contribute to content living multiple lives?” For the first life, we found that Facebook is the leading driver of giving content life with nearly 70% of the social actions. Google Plus came in second with 17%. StumbleUpon and Twitter came in 3rd and 4th place with 7.4% and 6.5%, respectively. For second lives, the Facebook and Google Plus still have the lead. However, StumbleUpon pushes its share to 7.3% while Twitter retreats to 2.2%.
Figure 8: First Life Social Drivers
Figure 9: Second Life Social Drivers
We know it is very difficult to produce good content. We also know that it is even more challenging to produce content that can withstand the test of time and resurface later when it is once again relevant (evergreen content). And while there still a lot more to explore in the many lives of content, our goal is to help publishers get a deeper understanding of their content and what works.
So what exactly are we trying to tell you? You can’t force a comeback, but you can certainly help once you see it take off. And if you see an article coming to life again, give it a push. It’s not often that you get a second chance.
About Wael Emara
Wael is a data scientist at SimpleReach. He is a data lover and machine learning/data mining enthusiast who thinks some mathematical models can read like poetry.View all posts by Wael Emara »
About Shane Jiang
Stat MS from School of Industrial&System Engineering, Georgia Tech. Data Scientist at SimpleReach.View all posts by Shane Jiang »