Social listening refers to the active monitoring of social media feeds in search of references to specific key terms.
You can track any term with social listening, but at TSV Analytics we like to group related key terms under a central theme, which typically corresponds to a company’s brand, industry or competitors.
While the concept is simple, the actual data returned through social listening is usually much messier than expected, and this happens for a variety of reasons. In essence, social listening casts a wide net and returns anything that gets caught, but the primary reason comes from a lack of context on the data collection side, which is elevated when keyterms have more than one specific meaning.
Acronyms, abbreviations, and homonyms are the primary culprits, but in many cases these words are required for covering the full scope of the overarching theme.
While in our context TSV means “Topic, Sentiment, and Virality”, in other places it could be referring to the common file type extension for “tab separated values”, or the airport code for Townsville, Australia.
So when searching for “TSV” through social listening, the wide net we cast returns posts referring to all of the examples above in addition to other uses outside of the context we’re looking for.
But because it’s our brand name, we’re not going to avoid tracking this term just because it has multiple meanings, otherwise we’d lose out on posts that might be directly referring to us.
The most straightforward approach would be to manually read through each post returned from social listening to better understand the context on a post-by-post basis, removing the ones that aren’t relevant.
This is easy to do when there aren’t many posts to look at, or you have unlimited time, but for a conversation with substantial volume this manual approach isn’t effective.
Instead of doing this classification manually, TSV Analytics uses AI and ML techniques to group similar posts together, creating groups of similar conversation posts, also known as topics.
Most of the heavy lifting here is done by a computer so that the user just has to quickly analyze each topic and decide if that topic is relevant to the overarching theme or not.
Posts from off-topic categories are then dropped, and any proceeding analysis can be done just using info from the relevant on-topic posts. Post-by-post determination can still be done at this stage for an even finer selection, but most of the work can be accomplished through topic selection on its own.
Rather than spending your day manually reading through posts to see what conversations are happening, test out TSV Analytics topic selection in your niche today!