Identifying Influential Content and Media

Understanding what content has organically been picked up and spread by an audience is a key insight to inform future creative campaigns.

Typically when looking to identify viral content you'll want to look at:

  • when there has been a peak in sharing activity by your audience.
  • which are the most popular pieces of content being shared.
  • which are the key sources which people are sharing content from.
  • how does the content shared and source vary by demographic group.
  • how quickly has each piece of content spread.
  • which are the personalities and topics driving sharing.

With these insights you can create more compelling, better targeted content in future that has greater impact on your target audience.

In this article we'll outline how you can build a solution using PYLON.

Identifying viral content using PYLON

As a quick reminder, you work with PYLON by:

  • Filtering the stream of data from Facebook to stories and engagements (such as likes and comments) you'd like to analyze. Filtered data is recorded into an index.
  • Classifying the data using your own custom rules to add extra metadata for your use case.
  • Analyzing the data you have recorded to the index.

Specifically when you look to identify viral content you will:

  • define an interaction filter that captures a broad set of stories and engagements on stories created by your audience.
  • define classification rules to identify topics, sources and personalities which influence content sharing.
  • analyze your recorded interactions to identify viral content being shared.

Solution overview

You may have seen a demonstration of PYLON that focuses on the movie industry. This makes a good illustrative example for explaining a solution.

Filtering interactions

Firstly you need to define an interaction filter that will capture a broad set of interactions for your target audience.

If you wanted to look for viral content influencing movie discussions you would look to capture a broad set of conversations and engagement relating to movies where a link is being shared.

You could use CSDL similar to the following:

    fb.topics.category in "Movie,Film,TV/Movie Award" 
    fb.parent.topics.category in "Movie,Film,TV/Movie Award" 
AND links.url exists

This example filter uses topic categories to filter to movie-related stories and engagements on these stories. The filter also states that a link must be shared in the story posted or being engaged with.

Topics give good coverage for well-known movies and are kept up-to-date in the topic graph. If you're investigating less well-known movies you should consider using keywords to match movies and plan to keep your filter up-to-date with current movies.

tip icon

Learn how to make use of super public samples to improve your filter by reading our guide.

Classifying topics and content sources

Before you start recording data from your filter you will want to add classification rules. Classification rules surface and normalize features of interactions for use in your analysis.

In this case topics provide excellent coverage of movie and personalities mentioned in stories. Depending on your use case topics may save you from creating a large number of classification rules.

That said, classification rules can always help you get more from your data. For this example you could add rules which classify the source of content being shared:

tag.webcontent "Social Networks" ​{ links.url any ",,"} 
tag.webcontent "Video" ​{ links.url any ",,,"} 
tag.webcontent "Music Streaming" ​{ links.url any ",,"} 
tag.webcontent "Shopping" ​{ links.url any ",,,"} 
tag.webcontent "Broadcasters" ​{ links.url any ",,"} 
tag.webcontent "Online News, Blogs" ​{ links.url any ","} 
tag.webcontent "News" ​{ links.url any ",,"} 
tag.webcontent "Gossip" ​{ links.url any ",,"} 
tag.webcontent "Fashion" ​{ links.url any ",,"} 
tag.webcontent "Sports" ​{ links.url any ",,"} 
tag.webcontent "Anime, Manga, Comic, Fun" ​{ links.url any ","} 
tag.webcontent "Review" ​{ links.url any ",,"}

These example tags illustrate how you can classify content source from the domain of links being shared.

Once you've finished defining your filter and tags you can start your recording. Interactions which relate to movies will be stored into your index. The interactions will be classified based on the domains in your tag rules.

When you come to defining tags for your real-world use case you will need to spend time iterating and testing these rules.

Example analysis results

When you have started to record classified interactions to an index you can analyze your recorded audience. Let's look at some typical analysis you may want to include.

Audience activity

You can use time series analysis to investigate when your audience has been most active.

For example this analysis between the 26th February and 2nd March shows peaks around the 29th February and 1st of March. Of course the 2016 Oscar ceremony took place on the 28th February. Analysis is based on UTC times and in fact the peaks coincide with the ceremony and the time of day US audiences are most active on Facebook.

You can 'zoom in' to the time series by specifying a shorter time window for your query:

You can see there is a peak at 3am (7pm pacific time) which was 1.5 hours into the ceremony. You can specify the same time window for your further analysis queries to focus on content being shared in this period.

Audience demographic composition

Before you look in detail at the content being engaged with it's worth looking at the demographic makeup of the audience that is engaging.

Using the and demographic targets in a nested query you can analyze the composition of the engaged audience:

The chart above show the composition of the audience, but to see how the audience compares to the wider Facebook audience you can use a technique called baselining. Redrawing the chart with a 'baseline' (shown in grey) shows how the audience compares to the general population of Facebook.

Here you can see that females between 18 and 24 engaged relatively less with the topic than they generally do on Facebook as a whole.

tip icon

Learn more about comparing audiences by reading our baselining analysis results pattern.

Viewing verbatim text

The above chart shows a peak of activity for on the 29th February. PYLON's privacy model prevents you from seeing the text of non-public posts, however you can take a look at super public samples during this period to get an understanding of the conversations taking place.

The pylon/sample endpoint gives you access to super public posts. An example post from this period read:

Finally Leonardo has got an oscar - about time too!

Looking further at a many super public posts during the period Leonardo DiCaprio is a strong theme, which is no surprise.

Popular content being shared

PYLON provides a number of analysis targets that allow you to inspect links shared by your audience.

You can use the links.url target to analyze content being shared the most:

You can segment your audience by using the target in query filters, and repeat the analysis for each group:

You can use nested queries, utilizing the and targets, to analyze popular content by demographic group:

As you'd expect in this case many of links relate to stories focused on the Oscars.

Popular sources of content

It is also important to take a look at which sites are popular sources of content.

You can carry out the same analysis as shown above but instead for domains by using the links.domain target:

You can also benefit from the classification rules added to your filter, in this case we classified sources of content. You can analyze the breakdown of sources of content using the interaction.tag_tree target, specifying the 'webcontent' tags:

Again by using nested queries, working with the and targets, you can analyze which sources of content are shared by which demographic groups:

Speed of content spread

Another interesting aspect of viral content is understanding the rate at which it was shared. Was the content shared by a large number of people for a short period? Or is it content that has gradually been shared over time?

You can investigate this aspect by performing time series analysis and specifying the link you'd like to investigate in the query filter.

This chart shows the result of analysis for two popular links. The chart shows the cumulative number of interactions that have shared the link over time. You can see both the overall number of shares (final height of the line) and rate of sharing (the gradient of the line).

Influential topics and personalities

Aside from the actual pieces of content being shared, you may want to look at the topics influencing conversations in your audience. For this example topics are useful as they provide excellent coverage of films and personalities.

By analyzing the time period using the fb.parent.topics.category_name target, it is straight-forward to show which topic categories and topics are prominent. In this case you can see that movies nominated for awards are trending:

Related design patterns

Take a look at these related design patterns to learn more.

Improving filters using super public data Remove noise and broaden the audience captured by your interaction filters
Baselining analysis results Learn how to baseline you results against a comparison audience