MIT Technology Review
A persona is an imaginary figure representing a segment of real people, and it is a communicative design technique aimed at better user understanding. For several decades, personas were data structures, static frameworks, user attributes without interactivity. A persona was a means of organizing data about the imaginary person and presenting information to decision makers. This wasn’t really workable in most situations.
How personas and data work together
With increasing analytics data, personas can now be generated using big data and algorithmic approaches. This integration of personas and analyzes offers effective possibilities to move personas from flat files of data presentation to interactive interfaces for analysis systems. These personas analysis systems offer both the empathic connection of personas and the rational insights of the analysis. With persona analysis systems, the persona is no longer a static, flat file. Instead, they are operating modes for accessing user data. The combination of personas and analytics also makes it less difficult to use user data for those who lack the skills or desire to work with complex analytics. Another benefit of persona analytics systems is that hundreds of data-driven personas can be created to reflect the various nuances of behavior and demographic in the underlying user population.
A “personas as interfaces” approach offers the advantages of personas and analysis systems and eliminates the respective deficiencies. Personas as interfaces change both the process of creating persona and analysis and offer, among other things, theoretical and practical effects on design, marketing, advertising, healthcare and human resources.
This approach of the persona as an interface is the basis of the persona analysis system Automatic Persona Generation (APG). To drive advances in the conception, development and use of persona and analytics, APG presents a multi-layered full-stack integration that offers three levels of user data presentation: (a) the conceptual persona, (b) the analytical metrics and (c) the basic data.
APG generates casts of personas representing the user population, with each segment having a persona. Based on regular data collection intervals, data-driven personas enrich the traditional persona with additional elements such as user loyalty, sentiment analysis and interesting topics requested by APG customers.
Using messaging system design concepts, APG identifies unique behavioral patterns of user interactions with products (i.e. these can be products, services, content, interface functions, etc.) and then assigns these unique patterns to demographic groups based on the strength of the association with the unique pattern. After we have a grouped interaction matrix, we apply matrix factorization or other algorithms to identify latent user interactions. Matrix factorization and related algorithms are particularly suitable for reducing the dimensionality of large data sets by recognizing latent factors.
How data-driven APG personas work
APG enriches the user segments generated by algorithms by adding a suitable name, picture, social media comments and associated demographic attributes (e.g. marital status, educational level, occupation, etc.) by querying the target group profiles of well-known social media platforms. APG has an in-house meta-tagged database of thousands of purchased copyright photos that are age, gender and ethnically appropriate. The system also has an internal database of hundreds of thousands of names that are also age, gender, and ethnically appropriate. For example, for an Indian woman in her twenties twenty years ago in India, APG automatically picks a popular name for women. The data-driven APG personas are then displayed to users from the organization via the interactive online system.
APG takes the basic user data acted on by the system algorithms and converts this data into information about users. This algorithmic processing output is actionable metrics and metrics for the user population (i.e., percentages, probabilities, weights, etc.) that are typically found in industry-standard analysis packages. Using these actionable metrics is the next level of abstraction from APG. The result is a persona analysis system that allows user insights to be presented at different levels of granularity, with the levels both integrated and appropriate for the task.
For example, C-level executives may want a general view of the users who personas would apply to. Farm managers may want a probabilistic view that the analysis would be appropriate for. The implementers must take direct user actions, e.g. B. for a marketing campaign for which the individual user data is better suited.
Each level of the APG can be divided as follows:
Conceptual level, personas. The highest level of abstraction, the conceptual level, is the group of personas that APG generates from the data using the method described above with a default value of ten personas. In theory, however, APG can generate as many personas as necessary. The persona has almost all of the typical attributes found in traditional flat file persona profiles. In APG, however, personas as interfaces enable a dramatically increased interactivity when using personas within organizations. The interactivity is provided so that the decision maker can change the default number to generate more or fewer personas, with the system currently set to five to 15 personas. The system can enable the search for a range of personas or the use of analytics to predict persona interests.
Analysis level: percentages, probabilities and weights. At the analysis level, APG personas act as interfaces to the underlying information and data that are used to create the personas. The specific information may vary slightly depending on the data source. However, the analytical level reflects the metrics and metrics generated from the basic user data and creates the personas. In APG, the personas provide the various analysis information via clickable symbols on the persona interface. For example, APG displays the percentage of the total user population that a particular person represents. These analytical insights are valuable to decision makers in determining the importance of designing or developing to a particular person and help address the issue of the person’s validity in representing actual users.
User level: individual data. With the help of the demographic metadata from the underlying factoring algorithm, decision-makers can access the specific user level (ie individually or aggregated) directly in APG. The numerical user data (in various forms) form the basis for the personas and analyzes.
The impact of data-driven personas
The conceptual shift of personas from flat files to personas as interfaces for a better user understanding opens up new possibilities for the interaction between decision-makers, personas and analyzes. With the help of data-driven personas, which are embedded as interfaces to analysis systems, decision-makers can, for example, provide analysis systems with the advantage of personas in order to create a psychological bond between stakeholders and user data through empathy and still have access to practical user numbers. There are several practical implications for managers and practitioners. Personas are now actionable because the personas accurately reflect the underlying user data. This full-stack implementation aspect was previously neither available for personas nor for analysis.
APG is a fully functional system deployed with real customer organizations. Please visit https://persona.qcri.org to see a demo.
This content was authored by the Qatar Computing Research Institute at Hamad Bin Khalifa University, a member of the Qatar Foundation. It was not authored by the editorial staff of MIT Technology Review.