Practically every episode of House M.D.‘s eight seasons had the same format: a patient comes down with some sort of mysterious illness, Dr. House and his team test, hypothesize, attempt to treat and, finally, accurately diagnose the ailment — all in less than 60 minutes! Usually, the team members diagnose the patient incorrectly at least once or twice before House swoops in with a genius idea the other team members haven’t considered yet or wrongly eliminated from the list of possibilities earlier in the episode.

The overwhelming “firehose” of social media data (think: THOUSANDS of tweets published every second) is a lot like the wide range of symptoms Dr. House’s patients usually exhibit: overwhelming and often misleading. How are social media marketers supposed to weed through all that content to find the information that’s relevant to their brand? With first-generation social media monitoring tools, it can be like hearing a list of symptoms and making an incorrect diagnosis without learning the context of the patient’s history and lifestyle: you can get metrics for your brand from social media, but often can’t see the context behind, for example, the number of positive or negative mentions.

Here’s a recent example:

Several days ago, on our sample social media stream for General Motors, we noticed an 86% spike in mentions for Cadillac - one of the search terms we’re tracking as part of the GM portfolio of brands – a confusing trend, since Cadillac hadn’t made any major announcements or hosted any major events that day. With a first-generation social media monitoring tool, we would have been left with “more questions than answers”, as a PriceWaterhouse Coopers report said earlier this year, even knowing the percentage of positive vs. negative mentions or the amount of mentions compared to an average day.

What we needed was to know the context behind that “symptom” of major spike in Cadillac Twitter mentions – what caused it? Why did it spread so quickly (the time frame of the spike happened between 9am and 2pm that day)? If we were managing the Cadillac brand’s social channels, we’d have wanted to know if it was something that needed immediate response, a response later on, or no response at all.

This is where we used Signals, a part of our uberVU product, to find the answers to these questions and accurately diagnose what the “illness”, or spike in brand mentions, was. Signals alerted us to the spike in the first place, where we would have had to wait for a retrospective data analysis to show the spike with another tool. It could have been another day or two before we noticed the trend! Like Dr. House and his team, the uberVU platform searched through a mountain of data points to find the root cause, and then pointed us directly to it:

A simple auto-correct error (ok, we’re REALLY giving @purpleheavy the benefit of the doubt on that one) turned the tweet into a source of an easily re-tweetable joke. Nothing for Cadillac to worry about, after all! In the world of Dr. House, this is one of those patients with a lot of obvious, deadly seeming symptoms that can be cured by something extremely simple. All in all, it took a big-yet-vague data point and solved it with a simple answer – all in 60 minutes or less :-)

 Tell us: how are you diagnosing what happens to your brand in social media? What else are you doing to uncover these insights and treat these social media “symptoms”?

Hugh Laurie image credit: Buou

Want to learn more about how to demystify your brand’s social media symptoms? Join us for a webinar with Forrester Research’s Zach Hofer-Shall on Tuesday, October 23rd at 2pm ET/11am PT where we’ll discuss how to Turn Social Media Into Actionable Business Insights.

Tagged as: Cadillac, General Motors, House MD, social media, social media monitoring, twitter, uberVU