Editor’s note: Jennifer Sikora is vice president of marketing at CivicScience, a Pittsburgh-based research firm.
As social media monitoring (SMM) gains increasing attention from marketing researchers and academics, much is being written about the data collected through it, both good and bad.
The reliance on SMM and reporting for consumer research insights is being criticized, and rightly so. SMM tools are good at tracking near real-time activity about subsets of social media users but these tools and practices paint an incomplete and often distorted picture and decision makers should pause before acting on the data.
An academic report published jointly by Carnegie Mellon University and McGill University (Science, November 28, 2014) discussed several limits to consider when using social media as a data resource:
- Inherent (“built-in”) biases are present in social media as a whole, particularly with certain sites.
- Social media sites use proprietary – and often undisclosed – algorithms to create or filter data streams. Those algorithms are subject to change without warning.
- Not all social media accounts are made by individual consumers. Some represent professional firms, bots, celebrities, consultants or even phantom accounts, while others are used for blogging or writing with a specific purpose.
- Many sites buy followers, so monitoring the data from these connections may not be objective.
Where does this leave SMM in the market research landscape? Well, it isn’t all bad news for social media data. Recognizing what SMM does well, while also clearly defining where its value drops off, is critical to leveraging it in the right way.
Use social media for the “what” and the “when”
Something happens. A brand launches a new ad or product. A new social marketing campaign kicks off. A PR crisis hits and the company begins to monitor social media for reaction. Is it positive? Negative? Can associations be made with other social media topics holding a particular sentiment? How many people care? Where are they? What else are they saying?
These questions are ideal for social media data to address – but SMM answers should usually not be considered final. Instead, reported data from SMM should be used to raise flags for marketers and researchers about where to dig deeper. At that point, other research strategies should be employed.
But can social media provide the who? Not really. Sure, you can see what the users’ names might be and grab a small amount of data about them through past responses or posts. But it doesn’t tell you how valid they are as a sample. The data also can’t provide deep comparisons between those who are responding very favorably to a new ad campaign and the general population. How these fans are different – in areas like shopping sentiment and behavior, personal finances and product adoption – can make a big difference to marketing strategy. Further, the data doesn’t tell you how active (or overactive) they are on social media in general. Are they reacting because it’s a genuine reaction or because they have a particular viewpoint that they are actively promoting? Is the person a troll who just seeks to stir the proverbial pot with their negative comments?
Reactions and feedback collected from social media can be swift and voluminous, particularly when large brands or higher-visibility topics are in the mix. The data can be collected but respondents can’t be probed further for additional explanation about their reactions (except in a very time-consuming, manual way, which would still only reach a very small percentage of users).
At this point other forms of research can be used to validate or further explore sentiment trends and to determine if vocal social reactions should even be considered. Two experiences come to mind that illustrate this practice:
Prior to a product line launch involving a celebrity licensing deal, the celebrity became involved in a negative PR scandal that received tremendous backlash on social media. Understandably concerned about the imminent product launch, the brand turned to consumer polls to gauge feelings about this celebrity’s bad press. While this would be difficult through SMM, a consumer poll was able to compare the brand’s fans to the general population. This data was able to show that fans of this brand (who would also be candidates for this product) were far less concerned about the celebrity’s scandal than the general population, giving the company data-backed confidence to move forward with the launch.
In another instance, a consumer products manufacturer decided at a corporate level to stand behind a social policy issue in its headquarters’ home state, resulting in a surge of negative sentiment toward the brand on social networking sites. Fortunately, the company did not recoil from its position as a result of the negativity on social media but continued to monitor its sales data, checked sentiment among fans and allowed the social reaction to settle down over time.
In both cases, social media data alone would not have been enough to make an informed business decision.
Social media opportunities in MR
One area of opportunity for social media data is in finding sentiment on lower-incidence products, brands or topics, due to the sheer size of the total pool of social network audiences. Finding a consumer research sample for say, an exclusive, small luxury brand or a new niche product can be very expensive and time-consuming to source using other research methods. SMM reporting can provide visibility to what’s being said about these lesser-known products, brands or subjects.
SMM and reporting tools and practices should continue to be used for fast and trended reporting on what is happening and when on social networks. This is what the tools are designed to do and they do it well. These metrics are part of most marketing dashboards today among consumer companies.
Most importantly, social media data research should be used to flag areas that need to be answered through other quantitative and qualitative research studies and not necessarily provide the answers to key questions. This will allow SMM data to be complemented with well-researched, in-depth and trended studies of audience profiles and allow for further segmentation and analysis. It’s a natural one-two punch combination that should make teams even more successful.