Listening: Should you filter social data for profanity?

I was at the Social Media Analytics Summit yesterday and was asked to fill in on a panel on one of my favorite topics: the challenges of social data. It was moderated by Nathan Gilliatt, and I was joined by Lisa Joy Rosner of NetBase (a client) and Catherine Van Zuylen of Attensity. We talked about the usual suspects: the challenges of sentiment analysis, definitions of popular metrics, that sort of thing. But the audience really perked up when we got to the topic of [cue dramatic music] profanity.

As customer advocates, marketers, analysts, executives, we’ve grown up expecting a certain degree of propriety and relevance from our data. It fits neatly into Excel spreadsheets and databases, drop-down menus and multiple-choice questions. We ask what we want to know, and we receive the generally appropriate response.

Social data, on the other hand, respects no boundaries. It displays us at our most human, in our casual, emotional, ungrammatical and sometimes wildly inappropriate glory.

So what should you do when you see a big, nasty swear dominating  your tag cloud or trending topics?

Our rational, logical side says “It’s data, and we have to pay attention to it so we can optimize.”

Our lizard brain says, “Run!!!”

Both Lisa Joy and Catherine have been there. They discussed the awkwardness that arises when they’re pulling together a report and there’s an unpleasant surprise lurking within. And they both told stories of clients who requested (even insisted) that they filter out the offending word or phrase.

It’s tempting. After all, we know there’s a lot out there that is spammy, irrelevant and a waste of everyone’s time.

With that, I will now tell you the story of the ass man.

Lisa Joy was in a meeting with a client, a shoe retailer. She unveiled the tag cloud, a compilation of the most common words used in the context of their brand. At the top right, in large, bolded type, was the offending phrase.

The client asked her to filter that phrase from the results. Lisa Joy suggested that before doing so it might make sense to drill down and see what it meant. The client was reluctant, but agreed.

As it happened, the phrase “ass man” was highly correlated with the word “slow,” and when they looked at the verbatims (examples of actual posts and tweets), they found a number of grumpy references to a “slow-ass man” (the store clerk), who apparently dawdled a bit too much for his customers’ taste when he went to get their sizes from the storeroom.

So the awkward, somewhat disturbing and apparently irrelevant phrase ended up yielding important insight about a customer service issue.

I can’t promise you that profanity will always have a silver lining, and there are certainly other considerations (Attensity asks employees to sign a waiver acknowledging that they may be exposed to profane language in social content, for example), but the lesson is this: resist the temptation to filter results before you have a sense of what they are.

And that, my friends, is the lesson of the ass man.

About susanetlinger

Industry Analyst at Altimeter Group
This entry was posted in Listening, Sentiment Analysis, Social Analytics, Social media, Social media measurement, VoC and tagged , , , , , , , . Bookmark the permalink.

2 Responses to Listening: Should you filter social data for profanity?

  1. ReaderX says:

    It’s pathetic someone would allow themselves to convinced by an attorney about any “need” to collect a waiver about being “exposed to profane language”.

    Like

    • Judge all you like, but IMO it’s an important disclosure, so people have the opportunity to make an educated decision about the content they are likely to work with as part of their jobs.

      Like

Leave a Reply

Fill in your details below or click an icon to log in:

WordPress.com Logo

You are commenting using your WordPress.com account. Log Out / Change )

Twitter picture

You are commenting using your Twitter account. Log Out / Change )

Facebook photo

You are commenting using your Facebook account. Log Out / Change )

Google+ photo

You are commenting using your Google+ account. Log Out / Change )

Connecting to %s