Introduction
One of the initial steps in any project is determining the target audience for the study. Social media research data has emerged as a primary source of information for companies. It has allowed the proliferation of users accessible as viable candidates for a specific domain. At one point, what was the job of specialized task forces to conduct surveys and find the appropriate user base?
It has now become much easier with websites facilitating research with the click of a button. Emails have normalized finding a target audience for surveys by collecting ad and user data from various websites. In such times, social media has played an enormous role in efficiently distributing requirements. Parallelly, there is a ubiquitous rise in user-sourced comments and interactions harnessed to gather insights.
Reasons for selecting Social Media data
The researchers primarily analyze comments to ascertain the needs and requirements of the research based on specific criteria. A few of the commonly known topics include:
- Sentiment: Positive and negative sentiment bias helps determine the popularity of a product.
- Employment: The marketplace’s reaction to a product or service is a good place. However, the employees of a given company can be seen as good pre-beta testers who have some baseline understanding of the product that the companies are creating. Instead of customized surveys, boards initiating discussion help build a better outcome.
- Typical user demographics: These can be targeted by specific age, location, and language factors. In the wide-scale usage of the internet, targeting a particular segment is often helpful as it restricts and better defines the scope of the research.
- Role and Popularity: Similar to the idea of applying a demographic base, restrictions based on specific expertise or position are also helpful. Influencers and social figures identified as ‘experts’ in a particular field are better suited to judge a product, topic, or situation.
Other than the segmentation, the selection of the comments is also important. Several comments are unfit for use and filtered out due to several factors, such as a lack of user information or unhelpful annotations. There are also generic or neutral comments, as well as those with a complicated vocabulary that is not quickly processed and unfit for use. As discussed in the book Social Media Analytics.
Problems with Social Media data
Social media algorithms are getting better equipped to handle these situations. They have means to select comments based on a black-box approach specific to a given platform. The trending feeds are better able to cater to the audience than the traditional newest posts. This has, however, also led to companies attempting to rig the algorithms. Biases introduced in algorithms lead to polarized distribution of content. Companies’ primary goal for their research and product or service development is user interaction, While it is still relatively early, they see mechanisms such as Machine Learning and AI as means to find better ways to improve.
