Group Created with Sketch.
Volume 20 No. 42
  • Created with Sketch.
  • Created with Sketch.
  • Created with Sketch.

Social media filtering grows sophisticated

Behind virtually every social media integration attempted by a sports team, league, university or network, there is a filtering component designed to keep the open, free-flowing fan discourse appropriate for mass audiences.

Beyond just looking for profanity, though, social media filtering has quickly grown into a sophisticated effort requiring constant vigilance to identify and screen out unsuitable words and phrases. When fans see social media feeds embedded into a live game video, arena or stadium scoreboard, or in other second-screen experiences, the filtering behind it is typically a combination of automated tools and human monitoring.

Properties must balance filters with the need to keep social media feeds valuable to fans.
Prior to starting a social media integration, the entity in question usually creates a long list of rules it wants. Within the list of rules, there are, of course, objectionable words, but also myriad combinations of letters and symbols, such as any team within a given league next to various spellings of “sucks” or “sux.”

“The lists are constantly updating. We’re adding strings of letters every day. More strings of letters, really, than specific words,” said Bob Bowman, MLB Advanced Media president and chief executive. “It’s sort of like vanity license plates. There is no limit to the creativity people have putting things together, and you have to stay a step ahead.”

And as social media and sports at large grow more global, the filtering also means having a better handle on foreign language slang.

“We don’t know every German swear word, but we’re getting there,” said Matt Corey, chief marketing officer for Mass Relevance, which aids numerous sports entities on curating their social media experiences. The firm screens social media in more than 60 different languages. “It’s all very sophisticated and has a lot of customizing.”

Computerized filters screen out posts with flagged words or phrases as they happen, but those automated tools are

frequently backed by staff members reviewing the feeds and acting as a safety net to identify problem posts not recognized by the computers. From there, other filtering methods can be attempted, such as screening all Twitter posts from users who have only an egg as their avatar, filtering along geographic lines, creating specific influencer groups, or screening in or out specific hashtags.

Photo sharing on social media, meanwhile, continues to explode in popularity. For now, filtering there is essentially all done on a manual basis, as tools do not yet exist to reliably identify and screen out objectionable elements in images on a fully automated basis.

Any filtering, of course, raises the issue of keeping social media feeds authentic and valuable to fans and not overly watering them down with heavy-handed controls. Both properties and vendors of filtering tools acknowledge it is a fine line to walk.

“You never, ever want to lose that voice of the fan,” Bowman said. “We have the filters, and you have to have them, but fans I think understand you’re not going to see a bad word in this type of setting.”