In spite of evidence that human work is increasingly social and relational in nature — and ample evidence to suggest that social and relational work is valuable to individuals and organizations alike — we have not, to date, built a language system and methods of measurement of the skills that allow us to assess whether people do such work competently or superlatively. And, as a result, we have not been able to develop ways to help humans demonstrably acquire these skills.

From LinkedIn knowledge graph reports to white papers of the World Economic Forum, from McKinsey & Co. and Boston Consulting Group to the International Labour Organization reports, soft skills have been touted as critically important to the current and next generation of workers and leaders. Moreover, demand for soft skills continues to outstrip supply — a large and increasing gap that has been confirmed by studies that probe into the main sources of dissatisfaction of chief learning and human resources officers with the skill base and skill development outcomes of training and development programs.

Why Soft Skills Matter

In every organization, soft skills show up as essential to both the vertical and the horizontal organization of work:

The vertical organization of work: In this model — per organizational economist Luis Garicano — organizations require social skills for conveying critical information, including decisions made and the rationale for them. Regardless of industry, organizations solve problems — whether it be client problems, problems related to their internal systems and processes or the removal of obstacles created by new government regulation. Within this class of problems, there is variation. Some problems are easy, others hard; some occur more frequently, others less so. Within the organization, there are workers who, over time, specialize in solving certain kinds of problems. Some specialize in solving the easy problems or those that occur frequently (let us call them “routine problem solvers”), while others specialize in solving problems that are exceptional because they are very hard, very rare — or both.

An efficient organizational form in such an environment turns out to be a pyramid-shaped hierarchy in which exceptional problems are passed upwards to managers or executives who either solve them themselves and communicate the solution to the routine problems solvers, or pass them upward to still more highly placed managers or executives. Because the latter have a comparative advantage in communicating solutions to unusual problems to the routine problem solvers, it is efficient for them to capitalize on this advantage and specialize accordingly, by acquiring social or communicative skills (soft skills) that allow them to excel in their roles.

This model notably assumes that the communicative prowess required to “pass an exceptional problem upwards” is less than that required to successfully communicate a solution to an exceptional problem to a routine problem solver, which is not always the case. But, the model explains the fact that “we have a lot of hierarchies around,” as well as the fact that “executives talk and write a lot” — and, perhaps more articulately than those the model would label as “routine production workers” in the same organization.

If we interpret the work of routine problem solvers as being largely cognitive and individualist in nature, and that of executives and managers as having significant social, interactive and communicative components and requiring skills that are not purely cognitive (i.e. “reading a room”), we glean some understanding of both why social and communicative skills are highly valued and why executives are so well paid (their skill set is sufficiently rare.)

This model is a variation on the classic command-and-control image of hierarchical organizations as devices that jointly minimize communication and coordination costs — in which information (about the environment) gets passed on upwards from employees through managers to executives, and decisions get passed on downwards (through the hierarchy).

In this model, being able to communicate decisions from above persuasively, eloquently and efficiently is a highly valuable skill.

The horizontal organization of work: This aspect of work requires convex combinations of hard and soft skills in any group or team, and in any member of that team. In the vertical organization model, most of the communication that matters is vertical: problems get passed upwards; solutions are communicated downwards. The model (correctly) predicts the growing importance of soft skills to managers and executives, but it does not explain why returns to soft skills have grown across both skilled and unskilled occupations, in both high-wage and low-wage sectors of the economy.

To account for this flaw, labour economist David Deming takes into account the horizontal structure of human work in organizations and argues that soft (or social) skills function as a sort of lubricant of social interactions in all settings in which humans work together, increasing marginal and average productivity. Given that more and more of human work is performed collectively and the quality of interactions among group members drives the group’s ability to solve problems, Deming’s “social lubricant model” of soft skills accounts for the uniform growth in return to such skills across wage and authority levels.

If every high-performing group needs to comprise individuals who have at least a certain level of communicative and interactive skill, then, in a collaborative economy in which most problems are solved in teams, there will be a shortage of interactive and communicative skills relative to an economy in which routine problem solvers are managed by interactionally sophisticated executives. This would seem to explain the growing value of such skills post-2000. Suppose, moreover, that with the increase in our collective, societal ability to select for, measure, train, certify and develop hard skills, the stock of certifiable hard skills has increased — which would lower their value, relative to that of soft skills.

Simply put, because soft skills are hard to measure and evaluate, it is hard to certify them in a way that lends credibility and market power to the owner of a “soft-skill certificate,” which would explain why the value of hard skills relative to that of social/interactive skills has decreased.

We can’t act upon that which we cannot see

Analyses of the demand for soft skills and the importance of soft skills to human and organizational performance reveal a perplexing multitude of different ways of defining and measuring them. Some define soft skills as “those skills that matter to lifetime outcomes and welfare but are not directly measured by standardized tests.” Labour economist James Heckman and his colleague Tim Kautz use the results of Big Five personality inventories (extraversion, agreeableness, openness, conscientiousness and neuroticism) to proxy for skills that traditional standardized tests of “purely cognitive ability” do not measure. They argue that individual performance and achievement depend not only on measurable cognitive abilities, but also on a host of other factors, some of which are stable (like personality traits) and some of which are situational (like incentives).

They find that long-run studies of individual study and work life average out situational factors and reveal the contribution of individual traits to life success. The intuition is that certain state- and context-independent characteristics — such as conscientiousness — affect the way in which someone responds to incentives and situational variables across different contexts in predictable ways.

Elsewhere, researchers base their findings on the U.S. Department of Labor O*Net database, which deploys a broad taxonomy of soft skills — on the surface unrelated to any of the Big five traits — and based on surveys filled out yearly by employers answering questions about the employment roles and positions in their organizations and the importance of various skills to the performance of the functions these positions entail.

Add to this the ever-growing collection of reports on skilling and development from advisory groups (McKinsey, BCG, Accenture, PricewaterhouseCoopers) and NGOs (UNESCO, UN), which use multiple incompatible definitions of soft skills and bundle together various sets of skills and label them as soft.

With such a wide dispersion of definitions — and, where applicable, methods of measurement — it is not surprising that soft skills have eluded targeted and constructive interventions aimed at either selecting rigorously for those who possess them or for helping humans develop them proactively.

…

Moreover, using the Big Five traits to measure the degree to which an individual possesses a skill so valuable sets up an implementation obstacle: Personality tests are based on self-reported answers to questions, which are asked in a way that makes it hard for the uninitiated subject to discern the link between her answer patterns and her reported trait.

In an era when not much rides on the results of such tests, there are lots of un-initiates around whose answer patterns are honest enough to produce the impressively informative correlations Heckman and Kautz present. But, if a large number of universities and employers suddenly asked applicants to take personality tests whose results would figure in to decisions made about their future lives, the incentives for cleverly distorting answer patterns to produce the impression of having the sought-after traits would create sufficient noise to vitiate the results of the test as a filter.

There is, clearly, a significant cognitive component in delivering a highly sensitive message to a divided or fragmented team, or in raising a knock-down challenge to a speaker in a public forum in a way that does not discourage her from responding in a truthful fashion. It is also easy to see that prototypically affective predispositions and abilities (“empathic accuracy”) can be instantiated in ways that are more or less cognitive: Just like one can be emotionally attuned and responsive to another person without being able to clearly and suggestively articulate the reasons for which one resonates with him, one can also clearly and persuasively articulate such reasons with a demeanour of such cool detachment as to lead others to doubt the genuineness of one’s empathic understanding.

X-factor skills that involve the deployment of executive functions of the mind/brain (i.e. problem decomposition into sub-problems, working memory use and capacity, impulse suppression, task switching) in different scenarios are in many cases labelled “soft.” They also condition the acquisition and exercise of hard skills: maintaining concentration on a programming assignment (impulse suppression); keeping all prior conditions of a theorem in mind while trying to prove it (working memory); and breaking up a hard problem into smaller, easier-to-solve sub-problems before proceeding (sub-problem decomposition).

These problems, alone or in combination, contribute to three persistent and counterproductive illusions regarding soft skills:

Illusion 1: Soft skills are transferrable

In other words, if you have them, you can apply them in different social and organizational contexts and knowledge domains. Unlike a mathematical modelling skill which is almost always about a particular topic and subject (a plasma, a waveguide, a truss structure), or a programming skill which may be both language- (C++) and algorithm- (convex optimization) dependent, a soft skill is not context-dependent. This illusion arises from the general and generic language in which these skills are often described (“helpfulness,” “creativity”), which makes it seem as if the domains to which they are applied and the contexts in which they are applied do not matter.

However, one can be very good at organizing, convening and structuring small groups of engineers and scientists and not be equally competent with small groups of musicians or very large, potentially virtual audiences comprising people who have vastly different backgrounds. One can be highly competent at communicating in precise, sensitive, attuned ways in small group settings and lose all sense of connection and responsiveness in front of very large groups.

Illusion 2: Soft skills are innate

By the time you enter the educational system, you either have them or you do not. There are genetic and epigenetic narratives around this belief, which, at various times, garner different levels of credence, but, one’s stock of soft skills remains largely unchanged by the time one enters the second or third grade of the K-12 educational system.

However, significant success has been reported in the development of skills that are almost unanimously labelled as soft, including: the ability to perceive and understand emotional landscapes and to self-regulate in the face of emotional challenges, relational abilities such as networking and negotiation and empathic understanding in emotionally complicated scenarios.

Illusion 3: Soft skills are irretrievably or inevitably soft

They cannot be measured, evaluated and trained in ways in which hard skills can because they always depend on some subjective and therefore necessarily biased report of an impression registered by a human observer subject to physiological and psychological vagaries.

However, until the design of tests that pinned down specific components of hard skills (such as the manipulation and solution of algebraic equations, matrix mechanics, theorem-proving procedures in real and complex analysis), hard skills were also subject to impressionistic biases, as individual examinations carried out by instructors produced results that could easily skew in favour of favoured pupils. Turning quality into quantity is hard, but inferring that something is impossible from the fact it is hard is not a useful step in this case. We can go from subjective evaluations to inter-subjective appraisals — which is the best definition we have for objective in any case.

The opportunity: Soft skills, hardened

We have an opportunity to become as precise about soft skills as we are about hard skills and create a set of measures, appraisals and methods for developing soft skills that match those we currently employ for the selection and development of hard skills.

The advantages of providing a satisfactory solution to the soft skills specification problem are clear:

We will be able to articulate, implement, explain and justify metrics and measures for a set of skills that are deemed highly important and valuable, but which are currently only recognizable after the fact — which for most programs and organizations means after incurring the cost of hiring and on-boarding an unskilled candidate, and the opportunity cost of losing out on a skilled one.

We will also be able to design training and development programs that make good on the now hollow and borderline fraudulent promise that many professional and other higher education programs make to help their graduates develop the skills they need for the jobs and roles they want.

A specification that cuts to the core of our intuitions and at the same time allows for operationalization using various measures at our disposal — subjective and intersubjective — is key to delivering on this opportunity.

If there is one core set of skills that almost all definitions of soft skills have in common, it is that they feature a strong social, relational and interactive component. If we can get to the core of sociality and relationality while retaining precision, we will have built a useful selection and development instrument for both organizations and education systems.

This article originally appeared in the fall 2024 issue of Rotman Management magazine and is lightly adapted from Mihnea Moldoveaunu's book Soft Skills: How to See, Measure and Build the Skills That Make Us Uniquely Human (De Gruyter, 2024). Subscribe now to unlock exclusive thought leadership and management research.