Every movement begins with a moment.
Think back to your last hotel stay. Whether you stayed at the Trump Tower or the Days Inn, when you arrived, you likely noticed that your room smelled of Fabuloso, the linens were clean and any traces of rodents — no matter how small — were nowhere to be found. You had your business meeting or enjoyed a few days on the beach with your family, and then you checked out. But did you actually take the time to fill out the survey questionnaire left on your bedside table? If you are like most people, chances are you did not. Why? The peak-end rule: People seem to perceive not the sum of an experience but the average of how it was at its peak (i.e., pleasant or unpleasant) and how it ended. You are more likely to fill out that survey if you experienced a very pleasant or a very unpleasant event during your stay. Needless to say, hotel survey results are highly skewed, as both very positive and very negative tails can produce bimodal distribution of data.
Phenomena such as the peak-end rule are not new to the psychological and research community. In fact, the list of biases and heuristics that psychologists and researchers must strive to eliminate from a statistically valid sample makes Atlas Shrugged seem like a tweet. As an analytics professional, I’ve always considered social media the equivalent of a highly biased focus group with an issue of observer dependency: The results obtained are influenced by the researcher. Did the unpaid intern monitoring your social media take a ride on the Google-sponsored Fur Bus last night? Has she successfully met the two-Red Bull threshold needed to power her sentiment analysis of your brand? Is she aware of statistical techniques necessary to analyze data that is not normally distributed? And what about the fact that hand-raiser and opinion biases make your social data nonrepresentative of your consumers?
I registered for and attended (most of) the “What Social Media Analytics Can't Tell You” session at SXSW. The description provided in the interactive schedule billed the session as, “Social media analytics can help you understand the active members of your social media audience, but what about the people who aren't posting? How do you fill in the gaps in your analytics with insight into your customers?” So, I was very much looking forward to hearing like-minded analytics professionals debate the best practices for dealing with data skews and biases. As I sat through the first 15 minutes of Alexandra Samuel, VP of Social Media, Vision Critical, discussing how the panel’s research questions were answered by combining social media analytics with survey data, my geek-chops started to salivate. Her first few slides confirmed what many of us in the industry already know: 68% of the online audience is silent — not represented via available social analytics tools — and they produce less than 25% of content. Her research concluded that the social media audience is a highly skewed proportion on the online population. I found myself thinking that this was the perfect setup to have that conversation I’ve been eagerly awaiting for the last 10 years of my career.
As Crowd Companies’ President, Jeremiah Owyang, took the stage to present the findings of a 90,000-person study of the collaborative economy, my expectations took a Jennifer Lawrence face-plant into the green conference center carpet. Why? Mr. Owyang started to extol his research findings. Online conversations can be classified in five groups: goods, services, transportation, space and money. Um, you had a 90,000-person study and you basically proved what people are talking about online by high, medium and low self-reported behavior? Don’t get me wrong, that’s a really great foundational finding — I just really expected this session to dig much deeper into how we as researchers and analysts can apply more scientific rigor around social media analytics to quantify behaviors, attitudes and perceptions.
I was already disappointed when Colby Flint, Consumer Insight Strategist, Discovery Communications, took the stage to discuss how social media users watch TV. Prior to joining Moxie, I worked for a large, broadcasting media conglomerate. My brilliant former co-workers produced several Advertising Research Foundation award-winning studies on exactly the same topic four years ago.
I couldn’t take it. I had to escape with my dignity and wanted the next 15 minutes of my life back to power up my mobile. Loosing mobile power at SXSW is social suicide, and I just could not justify jumping off that cliff for stale, outdated research.
Heather Carson is Director of Analytics at Moxie. Follow her musings on data, innovation and the Austin bar scene on Twitter: @TheData_Diva.
PLEASE PROVIDE YOUR INFORMATIONTO DOWNLOAD THE PAPER.