Going Green: Social Media Research
This project is a case study of St. Patrick’s Day and uses social media research to quantify unstructured text content for analysis and reporting.
A rigorous research process is outlined and followed with the aim of generalizing the research findings to a larger group, in this case online users. The study utilizes many online resources, such as blogs, forums, social networking sites, and websites to collect St. Patrick’s Day related content. This large amount of content is then reduced and sampled through random selection. The sample is then analyzed for sentiment and themes.
The following is an overview of the findings concerning St. Patrick’s Day:
- 2 out of 3 enjoy the holiday
- There’s not much anticipation for the holiday – most comments happen the day celebrated (68.3%)
- Businesses generate a significant amount of content related to St. Patrick’s Day (31.9%)
- The most discussed activities related to the holiday center on eating food and drinking alcohol (24.1%)
Social media research (SMR) taps into content that is voluntarily placed online. Blogs, forums, social networking sites (Facebook, LinkedIn, Twitter, etc.), along with website responses, and other online sources all possess the content of persons freely sharing their thoughts and knowledge with others. SMR is more than just “listening” to these comments. Social Media Research requires that an empirical scientific process be followed to gather and analyze social media data. Although there are websites that allow interested parties to view or listen to social media content, this ability does not constitute social media research (SMR). Understandably, there is a great urgency and desire to monitor the pulse of consumers, interpret their content, and make predictions. Without a reliable methodology and analytical plan, the findings and implications are not likely to have the desired effect or give answers to the questions you are asking.
For this whitepaper, a case study was performed examining the discussions and posts related to St. Patrick’s Day. The paper will quickly outline the steps used and the results obtained. Given the brief nature of this paper, a light subject is addressed. Dealing with complex social norms, behaviors, consumer preferences, etc. would necessitate more complex analysis. Because the purpose of this paper is to simply explore and highlight the process, advanced analytics were not performed, although it is worth mentioning that they can be performed using this research method for a deeper look at more complicated subject matter.
As mentioned, SMR, like traditional research methods (telephone, online, focus groups, etc.) requires an empirical process be rigorously followed. Much like traditional research, potential steps include: 1) description of the problem or goal; 2) theory and assumptions based on background information; 3) operationalization of definitions and measurements; 4) establishing the research design and methodology; 5) defining the sampling plan; 6) instrumentation for sampling desired population; 7) data analysis, and 8) recommendations, conclusions and interpretations.
The purpose of this study is to outline the SMR process through the examination of sentiment and content related to St. Patrick’s Day online chatter. Most discussions, it is assumed, will be positive and revolve around wearing green, food, and parties. To provide a good mix of content, a variety of sources are used including Rich Site Summary feeds (RSS), blogs, forums, social networking sites, etc.
The amount of online content that is constantly being generated concerning this subject and countless others is staggering. If company resources were not an issue, all content related to the subject at hand would be analyzed. Instead of drinking from the fire hose of online content, it is important that a random sample design be used to establish a more useable N size. The reduced sample size retains a sufficient amount of data and the findings can then be generalized to the wider online population.
SMR requires a different type of sampling frame. Instead of building sample based on respondent characteristics (financial advisors in Michigan, first time home buyers in rural Oregon, recently dropped cable services, etc.) as would be done in traditional modes, SMR builds its sample on website type (forums, blogs, social networks, photos and/or video sharing, and other types of websites where written content is accepted and available). As a result of the large amount of data collected from the various types of websites, a small but ample sized random selection was pulled from each source. In order to account for website penetration, since not all websites are visited equally, the data is weighted.
As expected, St. Patrick’s Day is seen as a light-hearted holiday meant for enjoying food, drink, wearing green, and time with friends and family. As shown in Chart 1, the majority of content was positive with only a very small negative percentage (67.1% and 3.8% respectively).
In Chart 2 we see that the largest social media content theme related to St. Patrick’s Day is actually associated to a business. Nearly a third of this content is neutral based, primarily offering readers an opportunity to spend time and/or money with the establishment (31.9%). Invitations that incorporate activities that mesh with the holiday’s traditions are likely to be more successful.
The secular part of St. Patrick’s Day tradition are a large part of social media content, such as wearing green, drinking alcohol, and time with friends and family. Only a handful of content was religious. The most mentioned activity is consumption of alcohol and food (24.1%).
The number of positive comments fluctuates over time, with the highest level of activity coming on St. Patrick’s Day (Chart 3). Most days preceding the holiday have very little content but combine for 32.0% of all content. The amount of content on St. Patrick’s Day compared to the preceding day increased significantly, 62.5 percentage points. Negative comments concerning the day did not appear until the day of the holiday. The drop in positive comments preceding St. Patrick’s Day is due to neutral content of businesses advertising.
In nearly all themes the sentiment is positive in nature and the enthusiasm is high (Map 1). The size of each theme directly corresponds to the number of times mentioned. The enthusiasm contained in each statement, regardless of being positive to negative, is also plotted. Persons posting well-wishes, generally with one or more exclamation marks, is both high in enthusiasm and sentiment. Wearing green was for the most part a positive sentiment, but not high on enthusiasm.
Conclusions & Recommendations
St. Patrick’s Day is a time of tradition -- traditions that are meant for personal and social enjoyment. As shown, each tradition has varying levels of enthusiasm, sentiment, and mention. With online individuals continually and freely providing details and information, social media research is a strong method for tapping the mind of the consumer quickly and inexpensively.
Use of social media as a research tool can be done as a standalone method or in conjunction with other traditional research methods. Listening to the consumer is only part of the equation. To ensure appropriate conclusions are made, the SMR process requires following proven research principles. Precision and an understanding of the online community are also vital components. Although quick-changing, much information about social media users’ mindsets on particular topics at a given point in time can be gleaned quickly and used to answer business questions in ways that give better understanding of the consumer, and can be used to drive business growth.