Edelman and Cision Form Communications Cloud Partnership Read More

How Statistics Can Improve Your Social Media Strategy

In a recent Wall Street Journal article (“What Celebrities Can Teach Companies About Social Media” published on October 14, 2015), “experts” offered some social media content advice:

  • From singer Rihanna – Cultivate platform-specific content for a multitude of social platforms.
  • From reality-TV star Kim Kardashian – Keep a perpetual conversation going.
  • From singer Taylor Swift – Interact with some of your followers and fans.
  • From actor Vin Diesel – Don’t be “overly promotional.”
  • From singer Beyoncé – Use social media to craft and perpetuate your message.

While it is entertaining to read about these celebrity social media accounts, best practices like these may not be helpful to many small and medium-size businesses. There are some statistical reasons that trying to emulate Rihanna or Vin Diesel may not be the best use of resources or even the most effective means for outreach.

What I want to do in this post is explain why you shouldn’t immediately rush to follow Vin Diesel’s social media blueprint, and also explain how you might pragmatically come to emulate Rihanna.

You are not a statistical outlier (and these celebrities are).

Imagine that you want to get really rich. You might consider talking to one of the wealthiest people in the world and asking them how to create wealth. The reason that this probably would not help you to create your own wealth is that two-thirds of the wealthiest people in the world either inherited or had a family with a substantial amount of money.

Which is not to say that they would not have anything to tell you about wealth, but they are extreme outliers on a bell curve of income distribution. The relevance of their advice might depend upon your proximity to them on the bell curve.

Similarly, celebrities like Rihanna, Kim Kardashian, Taylor Swift, Vin Diesel and Beyoncé are extreme outliers on social media. By their nature, social platforms have a power law (long tail) distribution. This simply means that for any random Facebook Page it is far more likely for you to have one fan than 100,000 fans.

In fact, a sample study of Twitter accounts a couple of years ago demonstrated this long tail distribution for Twitter, and also shows the chasm between the social status of these celebrities relative to most people and businesses on Twitter: the 99.9th percentile of Twitter users had 24,964 followers. T-Swizzle has 64.9 million.


Anytime you’re considering best practice data or advice from businesses that scale hundreds or thousands times bigger than you, you should question the relevance to your situation. There’s also another pragmatic statistical reason that you might be wary of creating your social content based upon celebrity best practice…

All populations are not equal.


You don’t need to look further than most infographics to find statistics whose datapoints contradict each other. Generally, this isn’t a computational error, it is an error of sample set.

Because we can’t sample every person in the world using Facebook, companies sample a population of Facebook users and generalize their datapoints to the entire set of Facebook users. Because our behavior on Facebook is hard for anyone except for Facebook to track, errors occur. Statistics are only as reliable as the sample population is representative of the whole, which is to say that most online statistics are wrong.

Let’s consider a hypothetical business on Facebook called Jim’s Plumbing. Jim’s Plumbing is a local Cincinnati-based company and has accrued over 3,000 Facebook fans in the three years that it has been an active social site, primarily building the fan base by incentivizing “Likes” with customer coupons. Any post that Jim’s plumbing sends out may reach 1-5 percent of his “fans:” 30-150 customers.

Compare that to Vin Diesel, the world’s third-highest paid actor in 2014. He has 95.7 million Facebook fans (up from 9 million in 2009). Each of Diesel’s posts reach about 950,000 fans, and he consistently gets at least 60,000 Likes for each of his posts (6 percent engagement for fans who see each post, .06 percent engagement overall).

Jim’s Plumbing would LOVE to emulate Vin Diesel’s Facebook presence, but is the population of Vin Diesel’s Facebook Page representative of Jim’s? Of course not. Vin Diesel’s fans gain social status from liking his Page and engaging with his posts. Because 9.88 percent of all Facebook users Like Vin Diesel’s Page there is a one in ten chance that this is a point of commonality between users. Social status is an important aspect of what we post and how we act on social networks.

Jim’s Plumbing’s fans like their Page to get deals and discounts. If they aren’t promoting their business on Facebook to the thirty or so people who see their content, they’re probably wasting their time there (spoiler alert: they may be wasting resources there anyhow). Fans of Jim’s Plumbing don’t enjoy any social status from engaging with Jim’s Facebook Page.

Point being, Vin Diesel’s fans aren’t representative of Jim’s Plumbing’s fans. They aren’t motivated by the same triggers to engage either site.

Looking for new ways to measure your PR efforts? Learn how with our recent white paper!

Statisticians start big (and so should you).

I don’t know how familiar most people are with Bayesian (statistical) analysis: FiveThirtyEight’s Nate Silver wrote a great book (The Signal and the Noise) describing his application of Bayes’ Theorem to political polls and sports, and there’s another great book by Sharon Bertsch McGrayne entitled The Theory That Would Not Die that describes the history of Bayes’ theorem and how it has been successfully used throughout the last century. Much modern statistical inference is done using Bayesian analysis.

You might ask what Bayes’ theorem has to do with social media communication and marketing. Hopefully a lot. Without getting into the computation of it, Bayes’ Theorem is a way of improving outcomes by calculating probabilities given additional information.

For example, you might want to determine the probability of a conversion event when you post on Facebook. A Bayesian would take a look at the probabilities of each event separately and the probability that you have a post on Facebook when you have a conversion event, and would determine the likelihood of Facebook posts resulting in conversion events.

That may be the worst explanation of Bayes ever, but what I want you to understand is that a statistician starts broadly and uses probability to hone in on the solution with the highest likelihood. And you should, too. You can’t start with a granular “Kim Kardashian” tactic and optimize your communication budget. You need to think like a statistician and begin broad. But where to start?

At about the same time that the Wall Street Journal published their piece, Pew Internet published a great article (with basic datasets) about social media usage among adults and different demographics.

As an example, here’s chart from the article:

Pew has started to include basic spreadsheet-style data in their charts, so you can upload these into a spreadsheet and do further manipulation or visualization with the data as demonstrated here by a tool I like called Chartblocks.

Here’s the same data from Pew in a different visualization:

And here’s a chart demonstrating the populations by age, calculated multiplying Pew’s percentages by Marketing Charts’ US population data:

Using bigger, more reliable datasets like Pew Internet’s social data can help to zero in on where your target audience is and what tactics would be best suited for your desired outcomes. Also, there are a lot of really cool tools to help you accomplish this.


What I wanted to demonstrate in this piece was how best practices in certain circumstances aren’t always applicable to other disciplines. While we might aspire to be as interesting, beautiful and popular as celebrities, our businesses most certainly are not.

Your resources will be better utilized by segmentation, targeting and by scrutinizing tools and platforms broadly first, than by starting from a granular perspective and hoping for success. That said, I want to conclude by wishing anyone reading at this point all of the social media success and popularity of Taylor Swift.


About Jim Dougherty

Jim Dougherty is a featured contributor to the Cision Blog and his own blog, leaderswest. His areas of interest include statistics, technology, and content marketing. When not writing, he is likely reading, running, playing guitar or being a dad. PRSA member. Find him on Twitter @jimdougherty.

Recent Posts

Cision Blogs

  • Communications Best Practices

    Get the latest updates on PR, communications and marketing best practices.

  • Cision Product News

    Keep up with everything Cision. Check here for the most current product news.

  • Executive Insights

    Thought leadership and communications strategy for the C-suite written by the C-suite.

  • Media Blog

    A blog for and about the media featuring trends, tips, tools, media moves and more.