Although we aim to develop a simple tool that every team can use, everything we do is driven by academic research and state of the art machine learning technology. We combine traditional behavioral research with big data analytics to uncover the patterns behind high-performing teams. Our algorithm relies on the following science:
Personality can be defined as the distinct patterns of behavior that characterize each individual. Personality is considered to arise from within the individual and remains fairly consistent throughout life.
Personality assessment is often used to help people learn more about themselves and their unique characteristics. Understanding your own personality traits as well as those of your teammates will help you leverage the individual strengths of each person of your team and support them in aspects they do not feel confident in. Making personality data visible to every team member will help you communicate better and create an environment where everyone can perform their best.
The Big 5 personality model became popular thanks to the research led by Paul Costa and Robert R. McCrae in the 1970s. Since then, the Big 5 personality model has become the leading framework used in virtually every scientific study of human personality. In addition to 5 personality traits (extraversion, openness, agreeableness, conscientiousness, and emotional range), research has identified 6 unique facets of each personality trait that are often more predictive of behavior in specific situations.
At Teamscope, we also rely on the research of professor Jüri Allik from the University of Tartu and his colleagues for the latest science in human personalities.
If personality best represents how we typically behave in a given situation, then values describe the motivational basis of our attitudes and behavior.
Teams that share similar values tend to be more cohesive, more engaged at work, and find it easier to reach consensus on important matters. Not surprisingly, candidates that share similar values to the team are up to four times more likely to stay with the company long-term.
We rely on the Schwartz theory of basic human values to understand what really motivates each individual. Developed by Shalom H. Schwartz, the theory defines values as intrinsic motivational goals that correspond to the following features:
We use the refined theory of basic human values (originally published by Shalom H. Schwartz in 2012) to analyze how values and motivation influence teams.
MyPersonality is probably the most extensive research project on human personality ever undertaken. The research, led by David Stillwell and Michael Kosinski from the University of Cambridge, gathered data from 7.5 million people, including detailed data from their social media profiles and scores from different psychometric and cognitive tests.
Their pioneering research uncovered a new opportunity for behavioral research – using social media data to understand the deep-level characteristics of people, with methods that show extremely beneficial scientific value. The algorithm developed by David Stillwell and Michael Kosinski, if given enough data from social media sources, is more accurate on average than a person's family members at predicting their personality characteristics.
Social media based assessment methods are quite new, so it’s not yet possible to rely completely and solely on them, however the algorithms are developing extremely rapidly and will offer a fast and convenient alternative to traditional assessment tools in the future.
Watson is the A.I driven platform developed by IBM that first became famous in 2010 when it beat Jeopardy champions on live TV. Since then, IBM Watson has developed into an AI platform that is powers medical research, customer service, risk management, education, financial service applications as well as many more.
At Teamscope, we rely on Watson to help us process natural languages – which means that Watson helps us examine the interactions between computer languages and natural (human) languages. Watson enables us to understand the sentiment and topic of each conversation, and even helps predict the personality characteristics and values of each individual. Incorporating Watson into our analysis allows us to understand people based on their blog posts, cover letters, resumes, and basically any other text they have written.
We combine all of the above mentioned methods for a single goal – to get a better understanding of the deep-level characteristics of each individual. Depending on the amount of data each user provides, we work with thousands of data points to predict 80 distinct individual characteristics. Then, our algorithm calculates all of the interactions in a team to understand the team culture and predict which characteristics to look for in a candidate.
We combined knowledge from roughly 200 academic studies to develop our data model. Team relationships are complex and there is no single theory or model we can rely on, but there are a lot of data points that can be used to predict specific aspects of team performance.
With this in mind, we have optimized our algorithm to look, among other data, for:
Similarity in values. Teams that share similar values tend to be more cohesive, engaged, and aligned towards a common goal, and candidates that share similar values to the team are up to four times more likely to stay for at least two years.
Diversity. Cognitively diverse teams are found to be more creative and are less likely to get stuck in their comfort zone. Diversity in education, experience, skills, and demographic characteristics is generally great, but in some personality characteristics (e.g. conscientiousness), too much diversity is found to create unconstructive tension in teams.
Sounds simple, but in reality it’s far more complex. Calculating the average values of these characteristics does not say much about the team - to really understand the team dynamics, we evaluate each interaction in a team and compare how each team member and candidate would influence the team as a whole. For a team of 7 people, we consider tens of thousands of data points.
In addition to the initial assessment, many of our client’s teams continue to use Teamscope long-term, which means we know how teams develop over time. Based on that feedback, we constantly refine our algorithm, so every team we analyze makes our predictions more accurate and helps us provide even more insight to the next user. By recommending Teamscope to your friends or colleagues, you are helping us help you build better teams.