By Sadulaeva Teona (trsadulaeva@edu.hse.ru)

Persona discovery is a the process of identifying and researching different types of users, who might engage with a product. Marketing strategies design is based on such personas, ensuring that the product reaches its target audience. Throughout the years, the process of narrowing down the silhouette of an end user has been done via thorough research and various polls for potential buyer. But lately there's been a tendency of simplifying analytic jobs with the help of big data and machine learning. One could say that these tools have become essential for persona discovery, providing insights into user behaviors and preferences.

One of the main benefits of using big data for persona discovery is the amount of available information. Companies gather data from a wide variety of sources, such as search engine queries, social media interactions and transactional data. Gathered data is analyzed to reveal patterns and trends, that can be used to differentiate groups of users. Moreover, big data also provides valuable insights into user behaviors, such as clickstream analytics, that capture users browsing habits. With this information, businesses can create user personas that are more accurate and specific than traditional methods that rely on surveys or focus groups.

Machine learning is the other important tool for persona discovery. These algorithms use data to identify patterns and make predictions on users future behavior. For example, a machine learning model can analyze groups history of purchases and predict, which users are most likely to buy a particular product. An additional benefit of using machine learning for persona discovery is that it allows for continuous learning and refinement. As the model collects more data, it becomes more accurate at identifying patterns and predicting user behavior. This continuous improvement means that businesses can update their personas as their target audience's needs and preferences evolve over time.

First disadvantage of said approach is the potential for bias and inaccuracies in the data. It's been established that data sets are often collected through automated systems and therefore can be prone to errors. Machine learning models are trained on this data, meaning any biases or inaccuracies in the data are being accumulated, which can lead to inaccurate or misleading persona profiles. For example, if the data includes an overrepresentation of a particular demographic group, the machine learning model may be biased towards that group.

Another potential disadvantage is the lack of human oversight and interpretation. Machine learning algorithms are designed to analyze large volumes of data and make predictions based on patterns, but they lack the ability to understand the nuances of human behavior. That is why the resulting personas can't represent users' motivations, emotions, and other factors that influence their behavior.

Final con of inhumane approach is the cost of implementing big data and machine learning technologies. Hardware and software required to collect, store, and analyze large volumes of data can be expensive, and the associated maintenance and support costs can be significant as well. Additionally, the cost of hiring data analysts and machine learning experts with the necessary expertise to use these technologies can be prohibitive for many businesses. Moreiver, there's a potential for businesses to become too reliant on persona discovery using big data and machine learning. While these tools can provide valuable insights, they should not be the sole source of information for marketing and customer service efforts. Businesses should continue to conduct research, interact with their customers, and gather feedback to ensure that their personas are accurate and up-to-date.

In summary, analyzing huge amounts of data, businesses can create detailed user personas that can be used to drive marketing and communication strategies. With the aid of machine learning algorithms, these personas can become more accurate over time, improving the effectiveness of marketing campaigns and providing valuable insights into user behavior.

It seems that while persona discovery using big data and machine learning can provide many advantages for businesses, there are also potential disadvantages to consider. The accuracy and bias of the data, the lack of human oversight, privacy concerns, cost, and reliance on these technologies are all potential challenges that businesses should be aware of as they consider using these tools. While machine learning and big data technologies provide a plenty of time-saving tools, persona discoveries demand human approach. Operating large chunks of data won't help with accumulated biased errors.

1. J. Salminen, B. J. Jansen, J. An, H. Kwak, and S.-G. Jung, ‘‘Are personas done? Evaluating their usefulness in the age of digital analytics,’’ Persona Stud., vol. 4, no. 2, pp. 47–65, Nov. 2018.

2. D. Cabrero, H. Winschiers-Theophilus, and H. Mendonca, ‘‘Usercreated personas—A micro-cultural lens into informal settlement’s youth life,’’ in Proc. Hum. Workshop Interact. Design Work Anal. Interact (Advances in Information and Communication Technology), J. A. Nocera, B. R. Barricelli, A. Lopes, P. Campos, and T. Clemmensen, Eds. London, U.K.: Springer, Jun. 2015, pp. 57–70.

3. https://www.tandfonline.com/doi/full/10.1080/10447318.2021.1908670

4. https://humansofdata.atlan.com/2022/10/improve-data-discovery-persona-driven-strategies/

5. https://blog.datagran.io/posts/find-your-buyer-persona-with-machine-learning

6. A. Przegalinska, L. Ciechanowski, A. Stroz, P. Gloor, and G. Mazurek, ‘‘In bot we trust: A new methodology of chatbot performance measures,’’ Bus. Horizons, vol. 62, no. 6, pp. 785–797, Nov. 2019.