In a lot of ways, the use of social data has not progressed much farther than meteorology a century ago. Oft-utilized metrics such as volume, sentiment and influence tell us what happened to a piece of content in the same way that barometric pressure, wind speed and temperature could tell us how the weather was developing for a limited area. Meteorology today has the ability to look globally at the macro patterns that come together, allowing us to forecast (however imperfectly) trends far in advance. Analogously, there are high-level factors which can affect how stories develop and move in the digital space as well.
An easy example of how this works is to use, with apologies, the weather. It turns out that discussion about weather events is actually more predictable than actual forecasts, and the Nor’Easters that pummeled the East Coast this year are a great example. Here is a media platform timeline of winter storm Juno, which barely missed New York and slammed into Boston, dumping more than two feet of snow:
Weather enthusiast and frequent flyer forums (dark blue line) lead early on, as they speculate about the impact and track of a storm as soon as it hits long term models. This is then followed by Twitter (light blue), where the hype eventually crosses over into a broader, non-enthusiast audience, with a limited amount of mainstream attention (pink line), usually from outlets like the Weather Channel and local broadcast stations. Then, the day that it hits, wire services carry the headline “Blizzard Slams Boston with Two Feet of Snow” far and wide. Interest then quickly fades as attention turns elsewhere. This exact pattern played out with each of the major Eastern snowstorms this year – unfortunately for Boston. Because the conversation adheres to the same behaviors, we can then identify common characteristics and what the factors are cause those variances. If you took over Old Man Winter’s job, and was the communications manager for a blizzard, here is what you would want to do to maximize coverage:
- Be first – Juno was the first major storm of the 2015 season and received double the volume of other winter storms that followed, regardless of accumulation of impact area.
- Location over impact – proximity to major metro areas, especially New York, was a major factor across all media types.
- Let them know you are coming – weather and travel forums will build early volume based on how quickly the long-range weather models come into general agreement.
All of these have a bigger effect in coverage than factors such as compelling visuals or accumulation amounts.
Similar patterns can be identified in any number of other situations, positive or negative. Hollywood films are usually as predictable as their plots when it comes to social media performance, allowing studios to adjust marketing months in advance. Though each film genre has its own set of indicators, they all have two things in common: Reviews don’t matter prior to opening and apathy is far more dangerous than negative word of mouth.
Conversely, the depressingly frequent plague of cyber criminals has created its own pattern. Unlike blizzards, mainstream news outlets usually carry the story first, and then it is adopted by Twitter, and then will fade away shortly after, absent a few elements that can keep a story alive longer. Prior to December 2013, the number of records compromised was usually the determining factor in how much attention a cybercrime received. The date is significant as it marks the Target breach, where 70 million customers had their information stolen, and the scale of those affected is now a secondary factor. Novelty is now the main driver – remarkability in who the target was (Ashley Madison), what the data contained (U.S. Office of Personnel Management) or the motive (Sony Pictures). Without a distinctive angle, coverage declines rapidly after announcement, as the 70 organizations that suffered major breaches the same month as Ashley Madison can attest.
None of this is to say that we will read tweets like tea leaves and see the future any time soon. Social media is vast, but has a finite attention economy that can be driven by a dizzying number of variables. Out of nowhere, a global news story can draw attention away from an issue, a celebrity can vault a minor incident onto the front page or a mascot can be incorporated into a meme overnight. While these wild cards make true prediction impossible (for now), they do not prevent us from better understanding and anticipating what conditions need to be present to amplify the positive and mitigating the negative.
* With apologies to Yogi Berra!