One of the better books I read last year was “Superforecasting, the art & science of prediction”, by Philip Tetlock and Dan Gardner.
Through very thorough longitudinal research they detected some people are consistently better forecasters than other people.
In the book they outline the main characteristics of superforecasters. “The strongest predictor of rising into the ranks of superforecasters is perpetual beta, the degree to which one is committed to belief updating and self-improvement. It is roughly three times as powerful a predictor as its closest rival, intelligence”. Another important finding of Tetlock and Gardner: if you combine superforecasters in a group and let them work together, they become even better.
The lessons of Telock and Gardner can be applied in business, and I certainly see good applications in the HR domain. It would be nice to find an organization that is willing to start some long-term research. Let’s focus the research on the area of talent detection and development. In many organizations talent talent detection is a top down process. The HR Talent Manager conducts an annual process, where managers are asked to classify their staff in various talent buckets. Global standard is the 9-grid, a 3×3 matrix with on one dimension performance and the other dimension potential (see example below).
The most promising group of people appear in box 3:3, high performance and high potential.
The output of the various managers is consolidated, and the result is discussed in a calibration meeting. “If your Jack is a star, my Trudy should be up there as well”. This process has several flaws. Very often managers do not have a very good view on the performance of the staff in their organization, especially when there is no regular contact. People who cleverly manoeuvre themselves in the surroundings of the managers who make the lists, have a better chance of being detected. Another problem is that for many positions it is very difficult to measure performance, so most performance ratings are the result of a subjective rating on a scale (often 5-point scale) of a manager.
Finding the HR Super Forecasters
The research I would like to conduct in a large organization could be roughly as follows.
- We ask a large group of people (as large as possible) to make predictions about the future of young graduates in the organization.
- All the participants are asked to put a change (as percentage between 0 and 100%) to the following event: “Graduate A, hired in 2015, will still work in our organization on January 1, 2021, and she/he will have been promoted at least to job level X”.
- If you do not know the graduate, you cannot answer the question for that person. If you get to know her or him in the cause of 2016 you can start answering the question.
- Participants are asked to go through the list of names every month, and they have the opportunity to adjust the percentages.
- The forecasting for the group of 2015 will end in December 2016. In January 2017 we start with the group of graduates hired in 2016. We repeat this five years.
- Graduates who leave the company, are removed from the list.
- January 1, 2021 is D-day for the first group. We can check who of the forecasters were right.
- In January 2026 we will stop phase one of the research. We have the data of five years of forecasting, and we will be able to detect the probably small group of HR super forecasters: the people who were the best in forecasting the success in this organization of the young graduates.
- Going forward the organization will expose this group to the young graduates during their first year of employment, and will rely on the judgement of of the super forecasters to select the right talent for fast-track career development programs.
I am sure the real professionals can find ways to improve to design. Please do! Organizations who are open to explore ways to detect the HR super forecasters, please contact me.