2017 paid search trends

Storytelling with Social Media Data

January 29, 2016

Christopher Spong
At NHI, we use social media data to tell a story as it relates to our clients’ needs and business goals. Learn how to tell your brand's data story.

At Collective Measures we are strong believers in the power of data. It influences everything we do and enables us to demonstrate the value of our work. Social media, like many of the digital strategies we execute against, provides a rich pool of data from which we can extract insights and prove success. We use social media data to learn about target audiences, test content, and understand fan behavior as it might relate to our clients’ specific needs and business goals.

Last Friday, January 22, Christopher Spong presented “Storytelling with Social Media Data” at the MinneAnalytics conference hosted by Medtronic. A major theme from the conference centered on moving beyond dashboards and beginning to tell data stories.

As an organization, we are shifting more and more to telling data stories and using the insights we glean to make actionable and impactful decisions.  To build our story using social media data, we need a few key elements:

  1. Characters to interact with and influence the story
  2. A plot to lay out the storyline
  3. Action that will drive the story along
  4. And an ending to wrap everything up

1. Gathering Social Media Data

In order to start crafting a social media story, we will need to gather the data that will color our narrative. Most of the social media data points and examples referenced will relate to organic (or unpaid) social media strategies, but the metrics and measurement examples can be adapted to all sorts of paid and unpaid media channels.

  • Performance Data: Taken from the channels themselves, performance data is export from insights or analytics tabs on the social platforms themselves. This data includes key metrics like impressions, engagements, and engagement rate.
  • Social Listening Data: Data provided by specific “social listening” tools. Provides metrics on brand mentions, hashtag usage, URL shares, and helps us to add context to our narrative.
  • Website Data:  Using Google Analytics or a comparable website measurement platform, can can measure social’s influence on other channels. Depending on goals and strategies, website data can be extremely powerful, i.e. with tagged links we can compare the actions of fans we have directed to the site with general referral visitors. We can also see how social is contributing to key onsite conversion goals.
  • Other Data: Finally where applicable, we can analyze other data sources like organic search, concurrent media campaigns, public relations efforts, and brand lift studies. This can help to fully flesh out the scene in which our story is taking place.

2. Defining the Characters 

social media data

When building a story using social media data, the cast of characters is made up of channels and content groups. To get “high-school English” with the whole thing, generally a character has a few elements:

  • A character usually exists before the story can begin. Each channel or content group should exist before entering the story. At Collective Measures, we would recommended having a channel strategy in place and a content strategy setup before trying to measure against goals.
  • A character reveals information about itself through action and dialogue. Channels and content groups reveal information about themselves based on performance and the key performance indicators we’re using to evaluate them.
  • A character has motivations or aspirations. The goals of our content groups and channels are the motivators, which help to define how the characters interact with the story.
  • A character also interacts with and affects the plot.

Social Media Character Example: Taco Bell

social media data

For this example story, we will analyze Taco Bell’s Facebook page as our main character. The data that is presented is not factual, but has been estimated for the sake of creating strong examples. Also note that Taco Bell has 10 Million+ followers, so the data points are larger than what most brands would typically see.

Taco Bell makes for a great character in our story as they see a lot of engagement, they focus on a few main content categories, and have a balance of paid and organic posts.

Character Development

To understand more about the character, or in this case the Taco Bell Facebook page, it is important to examine some characteristics that will help determine the role it plays in our story.

  • History: The history of this character is easily learned through the website about section and by eating at Taco Bell periodically.
  • Personality: The personality can be defined by looking at photos shared, reading post copy, and by watching brand videos.
  • Motivators: By looking at comments on a few posts, it is easy to read into the main motivation of this character: To eat Taco Bell.

3. Developing the Plot

The plot of the story we are going to tell is illustrated using the data we have captured. Characters (or channels themselves) drive the plot with their performance.

To fill in the plot, we need to look at key data points, in a few ways:

  1. A social measurement strategy helps to provide context into how data is being segmented.
  2. Priority KPIs are then determined to help show a channel’s impact on the overall story.
  3. A content framework is also created to help evaluate what type of content resonates best on each channel.

For instance, pretend that you find that your Facebook page did really well in December. You would want to provide reasoning why using a social measurement strategy and the KPIs that we will lay out to show impact on the storyline. We would also want to create a content framework, which helps us dissect content groupings and evaluate their performance, similar to measuring channel performance.

1. Social Measurement Strategy Recommendation

Utilize total and average metrics to gauge performance.  It’s important to look at social performance data as both totals and averages when evaluating performance. For instance, you may see a high number of total impressions and engagements one month because you posted 30 times, when in fact your page sees stronger average post performance when only posting 15 times per month.

A rolling 6-month average helps to benchmark performance over time while accounting for variables. There are a lot of variables that are hard to account for when looking at performance over time, try benchmarking over a rolling 6 month period. This period presents enough time to collect valuable data, makes it easier to account for changes in channels and strategies, and prevents changes in channels and pages from skewing performance measurement

Account for some variables, when analyzing social media performance:

  • Posting cadence – How many times you are posting per month, per week, or even per day
  • Fans/followers – How many followers you have now versus 6 months ago
  • Algorithm updates – Channel algorithm changes in the last few months
  • Other media campaigns – Whether or not media is running on other channels
  • Seasonality – And if your business usually experiences seasonality

2. Priority KPIs for Social Media DATA

Here are some of the key performance indicators (KPIs) that we look at when measuring social as a channel. Some of these are specific to organic social measurement, but similar KPIs would be used when measuring paid social.

social media data

Generally the metrics we look at fall into a few different measurement categories: Acquisition, activation and awareness.

The primary metrics are best used for benchmarking and evaluating success. Secondary metrics can then be dissected further to provide context or further explain a finding.

Because the goal of each post can vary, whether you are asking for comments, clicks to the site, or simply looking to reach a large audience, it is best to first look at primary metrics as a whole, then dive into secondary metrics.

3. Content Framework for Social Media DATA

In additional to examining and measuring performance on our channels as a whole, it is crucial to dig deeper into the types on content that resonate best with fans. We recommend grouping all posts into various Pillars and Themes as a means for creating a content framework.

social media data

To explain our approach to a content framework at a high level, content is broken down into Pillars and Themes. Pillar are groups of content that ladder up to business goals. Pillars generally don not change over time and every post should fall under a Pillar. Themes fall below Pillars and are more specific groups of content that live within one Pillar or another. Themes are more interchangeable and can be swapped in and out depending on business goals and during seasonal periods. Under-performing Themes can be eliminated from the posting strategy if they are not meeting content expectations.

4. Examining the Action

social media data 5

After our measurement strategy is aligned, it is important to look at specific events throughout the reporting period that impacted the story we are going to tell. This is the action – a grouping of events that affect the characters and plot.

Outlying posts, top performing Pillars and Themes, and anecdotes on fan comments are pieces of information that can make up the action. For example, an outlying, high-performing post, like a video that received far above average engagement and impressions, can be considered part of the action.

Aside from looking at performance by channel, by Pillar and by Theme, other ways you can segment the data to help define the Action include:

  • Performance by post time
  • By post type (like a photo vs a video vs a link vs text only vs a photo w/ a link)
  • By creative used

If we were to utilize a social listening tool to pull all social mentions of Taco Bell in the month of December, we might find an interesting theme that comes up quite frequently. More than anything else, fans want to see the return of the Beefy Crunch Burrito. So much so that there is even a Facebook page with over 35,000 fans all waiting for the return of their favorite Taco Bell menu item.

This information is incredibly powerful, and not lost on our main character. Convincing the product team to bring the item back onto the regular menu, then posting about its return, would likely generate numerous positive mentions about the brand and strong engagement.

5. Crafting the Ending

As we develop an ending to our story, we would start to build a report that illustrates the important information that will enable us to adjust our strategy and improve. If we can continue to take the metrics that are important to us, and increase them as time goes on, we are seeing success in improving our strategy.

Because there are nearly an infinite number of ways we can look at social data and because many different things can happen over the course of our reporting period, no two data stories should ever end the same way.

Example Ending

social media data

The exciting thing about social data is the number of ways in which we can tell a story. Evaluating performance can tell us a lot about our fans and our customers. Making content changes that positively affect a social channel can be replicated in another medium, such as our website or in a paid media campaign.

Uncovering that there is an entire movement to bring back Beefy Crunch Burritos can be communicated to the Taco Bell product team. Bringing back the Beefy Crunch Burrito could help with sales in a slow month or with brand sentiment during a tough PR situation.

Now Tell Your Data Story

To briefly recap, use all the social media data you can to illustrate the characters, plot and action within your data story.  Dive deep into performance metrics, website impact, social listening insights, and every data source at your disposal to uncover valuable information. And end each report with clear takeaways and strong recommendations.