Fifteen Years With The Harnessing of Tacit Knowledge and Employee Predictions

It was a sunny April afternoon in 2003 on the Cornell University campus. I was sitting outside the Johnson School of Business, glancing at the world’s leading hotel school, the Cornell School of Hotel Administration. It made me reflect on the many rich insights that hotel staff accumulate from their daily operational experiences. Hospitality employees are in continuous interactions with guests, colleagues, managers, and people from the industry with whom they meet and socialize. It came to me that many of these accumulated experiences remain unspoken and unshared. Furthermore, these employee experiences may develop into visionary mindsets about the future states of a firm: What if these visionary mindsets of employees influence firm performance over time?

The Tacit Dimension

This question led me to study the work of Michael Polanyi (1891-1976), a physical chemist who was one of the novel philosophers of sciences in the 20th century. In his book Personal Knowledge: Towards a Post Critical Philosophy (1966) [1], Polanyi refers to tacit knowledge. He then thoroughly presents the notion of tacit knowledge in his book, The Tacit Dimension [2]. As Polanyi argues, tacit knowledge includes inherited practices, implied values, and prejudgments. It is both conceptual and sensory information, and the formation of images to make sense of something. Tacit knowledge is scattered bits of knowledge, which are brought together to help form new models or theories. In Polanyi’s words “we can know more than we can tell” [2].

In the spring of 2004, Professor Einar Marnburg—my good friend, former mentor, and Research Dean at the University of Stavanger, Norway—came for a writing visit to Cornell University. We started to talk about operational measures with respect to the type of activities frontline employees accumulate tacit insights about, and how we can aggregate tacit knowledge. We were drawn to the work of George Katona (1901-1981), a Hungarian-born American psychologist and one of the first to advocate combined studies of economics and psychology.

Consumer Expectations

George Katona, author of The Powerful Consumer (1951) [3], was originally trained as a gestalt psychologist working on problems such as learning and memory. It was during the Second World War that he became interested in applying psychological principles to macroeconomics. He devised the measures of consumer expectations that eventually became the University of Michigan Consumer Sentiment Index. The survey collection of consumer expectations enabled him to predict the post-war boom in the United States, at a time when conventional econometric indicators were predicting a recession [4]. Katona wrote many books and journal articles advocating the development of economic psychology and fostering the general ideas of consumer confidence–foundational ideas for modern behavioral economics.

In 2005, I continued working on ideas about how to harness tacit information and expectations of employees by combining Polanyi and Katona’s work. I traveled to Singapore to interview management and frontline employees at the legendary Raffles Hotel in order to achieve insights into what constitutes outstanding hotel performance and what elements make Raffles Hotel one of the most sustainable and prestigious hotels in the world. These insights led me to continue my journey to Alice Springs, Australia, where I attended a tourism conference in the desert and presented my preliminary work on relevant measures for employee sensing in hotels. I had an inspiring time discussing and receiving feedback from hospitality experts’ evaluations of these measures.

Employee Sensing of Firm Performance

The work of Katona made me aware of the benefits of addressing consumers’ predictions about the future. In other words, an individual’s judgmental prediction is the ‘tangible’ product of a tacit knowledge accumulation [5]. From 2005 to 2009, I empirically tested the predictions of frontline employees versus those of managers in judgmental time series across hotels in Scandinavia. Together with Professor Sigbjørn Tveterås, who works with applied econometrics at the University of Stavanger, I worked on the analysis and found evidence that frontline employees can especially predict important hotel performance variables such as team performance, innovation, competitiveness, and guest satisfaction that are linked with short-term fluctuations in financial firm performance [6]. The work was evaluated as ground-breaking by the international PhD committee, but the thesis also opened up many implications about how to measure the internal workings of tacit knowledge in the minds of people. These implications later led me to investigate the differences in prediction measures in organizations when harnessing explicit versus tacit insights, where this tacit knowledge is more related to the sensing of non-observable, fuzzy measures, such as the sensing of firm dynamic capabilities [7].  

Validation Studies

From 2013 to 2016, a PhD candidate of mine, Carsten Lund Pedersen at Copenhagen Business School validated the idea that frontline employees accumulate information of strategic importance in the telecom industry. Carsten investigated the ability of call center employees versus customers in predicting firm performance, and its potential utilization in strategic issue management. Carsten’s study of more than 150,000 individual forecasts—one based on 13,531 survey responses–was subsequently compared to measures of actual firm performance. Carsten’s study confirmed that call center employees in the telecom industry accumulate insights about future customer satisfaction that are more precise than the sample of customer judgments [8].

From 2015, I started to work with a master student, Christian Blem Charity, on studies in Copenhagen Airports. Our focus was the identification of strategic issues for dynamic and effective predictions. The study concluded that there was a need for harnessing tacit insights more effectively by means of using prediction software.  

It was around this time that James Surowiecki’s book, The Wisdom of Crowds gained widespread interest. The book presents cases on the use of ideation crowdsourcing and software mechanisms to foster innovation in organizations. It was also around this time that the empirical results of the testing of corporate prediction markets at Google and Ford were published [9]. Basically, prediction markets operate like virtual stock markets. Google’s prediction system was designed by its own software engineers and lets employees bet on probable and observable outcomes with the purpose of determining the likelihood of future events such as: “Will a project be finished on time?” and “How many users will Gmail have?” [9].

Concurrently, another master student, Anne Sofie Lind, had initiated a master thesis project under my supervision. In order to test predictions globally in a multinational corporation, Anne Sofie aggregated 672 predictions in judgmental time series from ten different business units and found that employees in these hubs can predict short-term fluctuations in key performance indicators (KPI).  This additional real-world case study, set in a multinational organization, allowed us to estimate how key factors influence the absolute forecasting accuracy of crowds of employees and their relative forecasting accuracy compared to individual judges. The results support the theory of wisdom of crowds that diversity and expertise in employee crowds matter in making employees “wise” in their predictions [10].  

The Foundation of Collective Intelligence Unit and Testing of Smart Crowds

In the beginning of 2016, I founded Collective Intelligence Unit (CIU) at Copenhagen Business School to concentrate on grounded and empirical research of collective intelligence, mainly studied in judgmental time series. The definitions and stream of research on collective intelligence are many, and they are found in psychology, sociology, sociobiology, political science, economics, and computer and information science among other disciplines. Yet, the approach of Collective Intelligence Unit is interdisciplinary and tends to study collective intelligence as an emergent property from the synergies among crowd predictions, software-hardware, and independent aggregation mechanisms to produce just-in-time knowledge for better decisions. 

Recently, one of our CIU project teams, consisting of Julian Johannes Umbauh Jensen and Frederik Kjøller Larsen, accomplished a data collection on the assessment of smart crowds for the economy. We find indications that diversity and expertise in crowds matter in making crowds accurate in their predictions of household credit, debt, savings, and unemployment over time: 1, 3, 6 and 12 months ahead of time. These findings are also congruent with studies on crowds and diversity – namely, that identity-diverse groups can outperform homogeneous groups due to their greater functional diversity [10], and that a group of experts can consistently outperform crowds with less experienced individuals [11].

In 2017, I received a grant from the Danish Industry Foundation, along with my colleague Professor Torben Juul Andersen, to continue the investigation of crowd predictions in organizations. In this study, we apply dynamic crowd prediction software to harness tacit insights from employees in global hubs from 2018-2020. In 2018, it was a pleasure for Collective Intelligence Unit to host the first Nordic conference on crowd predictions. The conference featured both national and international speakers on collective intelligence and the predictive abilities of crowds. As one of the corporate partners in the project, the LEGO Group presented their interest in the study of crowd predictions on perceived product value. The conference also presented the latest tools for understanding how the wisdom of crowds can improve forecasting when dealing with risk and uncertainty.

The Future of Harnessing Tacit Knowledge

What will studies of tacit knowledge and leading predictions from employees look like in the future? At the Collective Intelligence Unit, we have observed that academia, public institutions, and the business community have just begun to take notice of collective intelligence. In a recent document by the European Parliamentary Research Service on the ten issues to watch for in 2019, the EPRS focuses on collective intelligence as an emerging field of research that looks into how AI can be combined with human intelligence. Just as the speed and scope of digitization, connectivity, big data, and artificial intelligence are now taking us “deep” into places and powers that we’ve never experienced before [12], so too will the combination of AI and human tacit, intuitive, and unconsciousness processing be the future foundation of collective intelligence, and it will bring organizations to new places that today are unimaginable.  


[1] M. Polanyi, Personal Knowledge: Towards A Post Critical Philosophy, Routledge & Kegan Paul Ltd, 1966.
[2] M. Polanyi, The Tacit Dimension, New York: Doubleday, 1966.
[3] G. Katona, The Powerful Consumer, New York: McGraw-Hill, 1961.
[4] G. Katona, Psychological Studies Of The American Economy, New York: McGraw-Hill, 1951.
[5] A. Bubic, D. v. Cramon and R. I. Schubotz, “Prediction, cognition and the brain,” Frontiers in Human Neuroscience, vol. 22, no. 4, 2010.
[6] C. A. Hallin, “Exploring the Strategic Impact of Service Employees’ Tacit Knowledge: The Development of an Indicator for Forecasting Economic Performance of Hotel Companies,” University of Stravanger, Vols. PhD Thesis UiS, No. 77, 2009.
[7] C. A. Hallin, T. J. Andersen and S. Tveterås, “Harnessing The Frontline Employee Sensing Capabilities For Decision Support,” Journal of Decision Support Systems, vol. 97, no. C, pp. 104-112, 2017.
[8] C. L. Pedersen, “Using the Collective Wisdom of Frontline Employees in Strategic Issue Management,” Frederiksberg: Copenhagen Business School [Phd], Vols. PhD Series, No. 33.2016, 2016.
[9] B. Cowgill and E. Zitzewitz, “Corporate Prediction Markets: Evidence from Google, Ford, and Firm X,” Review of Economic Studies, vol. 82, pp. 1309-1341, 2015.
[10] L. Hong and S. E. Page, “Groups of diverse problem solvers can outperform Groups of diverse problem solvers can outperform,” The National Academy of Sciences of the USA, vol. 101, no. 46, p. 16385–16389, 2004.
[11] D. V. Budescu and E. Chen, “Identifying Expertise to Extract the Wisdom of Crowds,” Management Science, vol. 61, no. 2, pp. 267-280, 2015.
[12] T. L. Friedman, Warning! Everything Is Going Deep: ‘The Age Of Surveillence Capitalism’, 2019.


Carina Antonia Hallin, Head of Collective Intelligence Unit, Copenhagen Business School

Share this:

Leave a Reply

Your email address will not be published.