Mechanical Turk (MTurk) is a marketplace for research on issues that require analyses conducted with the help of human intelligence. The platform allows businesses to access a skilled workforce and gives an opportunity to perform a great variety of tasks, including sentiment analysis, categorization, and other (Amazon Mechanical Turk, 2018). After a review of MTurk’s HITs, it became apparent that the majority of recent requests are related to data collection and transcription.
For instance, some requesters asked workers to type texts from images, extract data from receipts, categorize variables, and find links between them. The primary advantage of the service is the possibility to recruit study participants/ workers from a large pool of individuals from highly diverse backgrounds. Secondly, Follmer, Sperling, and Suen (2017) state that it is “relatively dominant in terms of data quality and the number of studies that use it compared with alternative platforms” (p. 330).
The reliability of data gathered from MTurk is verified through several studies that implemented test-retest instruments (Follmer et al., 2017). Lastly, it allows obtaining a large volume of data at a low cost and in a relatively short time.
When conducting the Starbucks marketing research aimed to predict the demand for new drinks (for instance, banana milk frappuccino vs. vegan mango latte) through MTurk, the marketers can seek to explore which one of the customers would like more and what would be the motivations behind their choices. The main questions can be as follows: Which of these products would you prefer to purchase? Why would you prefer to purchase it (taste, ingredients, cost, or other reasons)? The first question can provide straightforward answers linked to study objectives, yet the second question is even more important because it allows identifying the variables that influence customers’ decision-making (Attest, 2017). This information can help improve current products and develop new, better ones based on consumer preferences.
To explore the first question, respondents will be provided with brief descriptions of products and sample pictures of drinks and asked to choose one of the two reply options. In this way, quantitative data will be obtained that will be easily transferred into percentages. The second survey item will include multiple-choice options, including an opportunity to provide an open text answer regarding product traits that respondents would deem particularly valuable. Thus, it will procure both quantitative and qualitative data that will deepen insight into the matter.
As such, there is no need to target the research on a specific population group because potential Starbucks customers come from different walks of life. Thus, a randomized (probability) sampling will ensure that their interests are considered and will also entail a higher level of finding generalizability. However, it is possible to analyze participants’ preferences based on their reported demographic data (age, gender, educational level, and so forth) that may be gathered through additional questions. Also, it can be useful to limit research to a certain region where the product will be launched because consumer behaviors may differ in distinct countries.
References
Amazon Mechanical Turk. (2018). FAQs. Web.
Attest. (2017). 6 simple ways to test consumer preferences. Web.
Follmer, D. J., Sperling, R. A., & Suen, H. K. (2017). The role of MTurk in education research: Advantages, issues, and future directions. Educational Researcher, 46(6), 329-334.