What is the best way to use and interpret data generated by social media monitoring?

Data from social media monitoring is a potential treasure trove, but has to be analyzed correctly.

Ray Martino
, Partner, head of PR, Martino Flynn

Pamela Meek, Senior director of external communications, SAS

Jamie Robinson, Research and insight director, We Are Social

Annelies Verhaeghe, Head of research innovation, InSites Consulting 

Terry Villines, VP, director of analysis, PhaseOne

Ray Martino, partner, head of PR, Martino Flynn
Aggregating social data is the easy part, but finding actionable items - now that is where things get tricky.

First off, rely on humans and not just technology. No social media monitoring platform available has perfected data analysis. Identifying influencers is heavily weighted by number of followers, retweets, and Klout scores; sentiment analysis is incapable of detecting conversational intricacies such as sarcasm; and most importantly, those that do offer advanced analytics have imperfect algorithms, so it requires human interpretation to double-check results.

While monitoring tools work for aggregation, demographic analysis, and key- word usage, it takes human interpretation to uncover the most valuable bits of data.

Secondly, move beyond volume and sentiment. Those are fine starting points, but properly interpreting data means you must be able to act on your analyses. If you're focusing on share of voice, identify the topics people discuss in relation to you and your competitors.

For example, a smartphone firm might monitor conversations for data plans, mobile broadband, Wi-Fi, and more. Once the topics are defined, conversations must be analyzed to gauge your company's next move. If people are trashing your product or service, track what they are saying. Taking note of dissatisfaction will provide guidance on how to improve your business. Consider it market research in real time.

Third, look to the future. Along with monitoring brand and product mentions, you should also track key words and phrases relevant to your industry as a whole. Tracking these discussions and picking apart the details can foster stronger collaboration between the marcomms and research and development teams, and inspire your company to be more forward thinking.

Pamela Meek, senior director of external communications, SAS
If you are engaging with customers in social media, you are sitting on a gold mine of information about how they feel about your brand, products, service, and people.

Social media analytics - the key to interpreting what's in the data - is foundational for reputation monitoring and even predicting what might happen next. You don't need a degree in analytics, but you do need to identify goals for what to monitor and apply appropriate analytics against those goals.

Text-analytics software monitors social media channels, filters out noise, and analyzes conversations pertinent to your brand for baselines, trends, and triggers. Sentiment analysis can identify brand advocates or potential problems - especially important for averting or managing a crisis.

But don't just analyze social media in a vacuum. The lines between social media, traditional media, advertising, and customer service have blurred to the extent that everything is relevant to PR.

Fashion retailer Chico's uses social media analytics to gauge public response to TV ads, modifying media placement and strategies accordingly. Automakers monitor call-center complaints alongside social media to predict parts failure before problems escalate. These businesses analyze social media and data in the context of organizational objectives. This kind of information grabs the attention of executives.

Social media data enables us to benchmark, set goals, and demonstrate our progress. Beyond a day-to-day snapshot, effective social media analysis measures trends and patterns. Is your share of voice increasing? Are messages reaching key audiences? Is brand awareness increasing among key influencers? Is your social media program resonating within targeted channels?

Jamie Robinson, research and insight director, We Are Social
There's a lot of talk at the moment about big data and the challenges and opportunities this data offers businesses. When people talk about this topic, they often focus on the big numbers; 80,000 tweets per minute during Usain Bolt's 200-meter Olympic race; 52 million fans on the Coca-Cola Facebook page; or 100 million Weibo tweets per day.

However, it's much more useful to think about social media's big data in terms of conversations. Conversations are where the true insight can be found, for example knowing what people are talking about or why they're engaging with brands, products, or services. For me, the real challenge is categorizing and analyzing an ever-growing number of conversations across several platforms in multiple languages.

Fundamentally, people drive insights, not machines. While this may not be the sexiest, headline-grabbing insight you expected, it holds true again and again, no matter the scale of the project or the volume of conversations. Structured into an ongoing research framework, this three-phased process has proven to work best:

  • Use social media technology to extract broad trends. Use keywords, natural language processing, or text-analysis software.
  • Use the technology to identify the most relevant conversations. For a major brand generating thousands of daily conversations, this may be to identify those that are only about a specific subject, and perhaps, limited to a specific time frame.
  • This is where the true insight is found. Here, experienced social media analysts will dissect all or a sample of those conversations to extract insights and learnings.

Think about conversations, not big numbers. Invest in social media technology, but more in people.

Annelies Verhaeghe, head of research innovation, InSites Consulting 
Data has no value without context. Netnography today is focusing on finding new ways of turning information into insights.

Technology has really helped us when it comes to analyzing online conversation. Text analytics has proven to be an added value. However, it is not only technology that can help us in finding the golden eggs. We should go beyond data and focus on context.

  • Understand the conversation context. Too often we focus on them in isolation. However, integration with other types of information can help us to prioritize. For example, many clients use social media netnography to investigate the health of their digital platform. It is a must in this type of research to also link online conversations to behavioral measures, such as number of likes or number of visits to the pages.
  • Understand the client context. If we want new and fresh insights from this huge amount of data, we need to know what our client already knows. Through in-company workshops upfront, we should connect the dots with previous research and identify those knowledge gaps within the company. By mapping what is already known, we can focus on conversations with fresh content.
  • Understand consumer interpretations. The future of analyzing user-generated content is not to measure what is already there. It is to measure what this information does with consumers. One can wonder what the point is of measuring the number of brand mentions on a page if this brand was not even noticed by the consumer.

Synergies between natural and research communities, which will allow us to understand the why behind conversations, need to be created. In a research community we ask consumers to select the most powerful online conversations. With them, we then research the drivers behind consumer activation.

Terry Villines, VP, director of analysis, PhaseOne 
Today there is a premium on real-time analysis of social media with many data aggregators striving to provide real-time metrics for activity in this volatile environment. While the appeal of having instant knowledge of mass behavior is certainly understandable, a more thoughtful approach will lead to significantly better results.

Real-time, aggregated analysis carries two fatal flaws, the first of which is false positives. Just because the crowd is tweeting about it doesn't mean they always bring those tweets into the real world. Analysis in real time can be misleading. For example, I'm excited about the new Tesla, a new all-electric car. While my excitement might cause me to aspire to have one, it probably won't lead me to buy one within the next year.

The second flaw is first false impressions. A person's first impression, the one tweeted because of excitement, frequently is not a lasting one. Once more data points are gathered, perceptions evolve. The considered, lasting reaction is most meaningful.

For example, I was wildly excited about the iPhone 4S, but once I had experienced it, I realized it was not substantially better than my iPhone 4 and I chose not to buy it.

Even with these flaws, are you harnessing social media monitoring for its full potential? Are you seeing evidence that your brand personality or brand promise is becoming part of the public's vernacular?

We know brands that embody a consumer truth and express it across platforms, including traditional outlets, are those that are successful at driving social media engagement. Even with all the listening tools we have, we can't forget our own role as communicators in the social media conversation.

The Takeaway
  • Data from social media monitoring is a potential treasure trove, but has to be analyzed correctly.
  • Don't ignore the human components in the data. Make sure your staff can examine potential trends with their own eyes.
  • Focus on actionable data and be proactive in leveraging what you learn from social media monitoring to tweak your messaging.

Have you registered with us yet?

Register now to enjoy more articles and free email bulletins

Already registered?
Sign in