We recently completed the first phase of a yearlong social listening project for one of our clients. It’s a global project, spanning seven countries and includes “listening in” on all of the popular social media channels – Twitter, Facebook, Sina Wiebo (a hybrid of Twitter and Facebook in China), YouTube, and countless blogs and forums.
If you’re just counting Twitter, Facebook, and Sina Weibo alone, that’s roughly 1.7 billion active users! So you might think, “Wow, with all of those users and conversations going on, you’ve probably got a lot of data to work with, right?” Well, surprisingly, there wasn’t. Out of all of the “big data” conversations you might have expected one would find in today’s extroverted world of non-stop micro-blogging, stream of consciousness-dumping, Facebook attention-getting activity, we were left with only 629 relevant conversations. And of the 629, just 36 were brand mentions. So what are some of the lessons learned?
Don’t always expect to get a large sample size
Before delving into any social listening exercise, don’t set yourself up thinking that the project will be a success only if you get huge volumes of clean data. The reality is that social data is messy, and after separating the wheat from the chaff, you might not be left with much to work with. Herein lies an opportunity to emphasize the stories that can be told from the data, rather than worrying that your sample is not representative. The value in social data comes from the qualitative learnings it can bring to your project. Keep your expectations reasonable so that you don’t get frustrated when you’re not able to measure the data it in a precise, statistical sense.
Don’t get hung up on measuring brand mentions
When it comes to prescription drugs, especially those used to treat serious or chronic conditions, most people don’t talk a lot about brands, and even when they do, they show no loyalty. This isn’t necessarily a sign that there’s something wrong with your brand, it’s just that the majority of people is not going online and saying things like, “I really love brand ABC over brand XYZ for treating my cancer…”
When patients and caregivers are online, they’re focused on looking for information and sharing the story of their journey about the challenges that come with treating their condition, rather than sounding off about the positives or negatives of a particular brand.
Be wary of automated sentiment analysis
Getting sentiment right is tricky even if it’s a human being who is doing the analysis. For example, if someone says, “Oh great, my doctor said that I get to start my chemo treatment on Monday,” without providing any other context, a human being might sense a sarcastic tone and classify that statement as negative, even though there’s a chance that it might be positive. Therefore, it’s easy to see why many of today’s automated sentiment analysis tools don’t always give you the results you expect. Don’t be afraid to dig into the data and challenge your vendor’s claim that they have the most accurate sentiment engine in the industry.
Focus on what you can do with the social data you have collected
Instead of obsessing over what you didn’t find, concentrate on the useful aspects of the data. Social listening can help you to:
· Tailor your communications – Analyze what you have, and use it to create messages that will strike a chord with your target audience and elicit a positive response.
· Influence media planning and buying – Relevant conversations can point you to new, previously unknown sites and channels that can be added to your media budget.
· Discover top sources – The few, key people who contribute a lot the conversation will be easily identifiable. Social listening lets you find those people and engage them in your marketing efforts as ambassadors or in other patient-to-patient roles.
The point here is to understand that a lack of social data doesn’t mean that your project was done in vain. Even limited data provides plenty of opportunities to fulfill business objectives through strategically-built listening efforts. In other words, lack of volume doesn’t equate to lack of insight.