With the growing excitement about a skills-based economy – skills-based talent strategies and skills-based technologies – we thought it would be useful to provide some context with which to consider the benefits and limitations of these new tools and strategies. More importantly, offer guidance for decision-makers about how to get the most out of the emerging tools while mitigating their risks.
It’s an exciting time for talent leaders and for all of us workers, as technology seems closer and closer to describing us head-to-toe. Perhaps in the not too distant future, software will be able to describe each of us in sufficient detail to predict our perfect job and find that job for us. Likewise on the demand side, it will describe the perfect candidate and find them for us. Perhaps it will help us redefine the work itself to be optimized for a pool of candidates made more perfect by redesigned jobs. Lesser versions of each of these use cases exist today, and the distance between where we are and this idealized future is partly a factor of how we define “skills” and grapple with the limitations of current models. In this article we’ll unpack these concepts in the talent tech arena, talk about the technologies available today and offer a set of recommendations for talent leaders deciding how to proceed.
How we define skills is critical to this conversation. As humans and talent leaders, we think of skills broadly as categories of tasks a person can do, as observed in the real world. The software that appears to know our skills is a long way from observing, parsing and evaluating real life human behavior. Instead, it is processing data from many sources – employer job histories, LinkedIn profiles, formal assessments, self-reported skills questionnaires, GitHub and elsewhere – and inferring a set of “tags” for each of us. These tags are generally words or phrases, often lacking behavioral descriptions, proficiency levels or calibration. In many cases, they are predicting types of work you may have been exposed to, as opposed to areas of mastery. The software then uses these tags, labeled as skills, in a variety of ways to make other predictions, like matching candidates with opportunities. There are significant potential benefits from this new approach to talent, but it’s worth keeping in our minds a distinction between the skills themselves and the data that attempts to describe them, lest we over-inflate the “intelligence” of the technology and diminish the human judgment we apply to our vital talent processes.
Despite the data quality risks, most talent acquisition leaders have been effectively using some form of skills filtering in their sourcing activities for some time, and internal talent leaders would be remiss for not at least considering the emerging Talent Marketplace tools. This new category serves as a means to create greater employee mobility and surface hidden talent from deeper in organizations than has been possible before. In addition to improving traditional talent outcomes like internal fill percentage, time-to-fill and retention, several recent case studies have shown “found capacity” by using their talent marketplace to staff projects employees contribute to on top of their regular jobs.
So, how get the most out of your talent data and tools in the new world of skills? Below are five recommendations to help you navigate:
- Recognize skills as a new, informal type of data.
Understand that the current operative definition of skills is far removed from the scientific approach HR’s competency nerds have long been working to perfect. This doesn’t invalidate the benefits of the new model, but it requires process design that takes into account the uncertainty of the data, ensuring some type of validation or dialogue occurs before actual talent decisions are made. Probe the specific methods your tech providers use to gather and infer skills tags, both behind the scenes in their machine learning models and at the user experience layer where much of the data is created and “validated”. Then design processes that seek to enhance the validity of the software’s predictions and refine them through more human-driven processes. - Apply a surgical approach to competency modeling.
Apply true competency modeling science where it matters. If you have clear technical tracks and the ability to assess KSAs (knowledge, skills and abilities) objectively, with clear benefits to doing so, roll up your sleeves to maintain these taxonomies and design them into the configuration of your talent tools. This level of rigor is best applied surgically to those segments of your workforce where the accumulation of specific skills is critical to developing or delivering your products and services for customers. - Match your organizational culture with the tools you deploy. Talent marketplaces are designed to create transparency and a healthy churn of talent. Organizations with more paternalistic succession planning or where cross-pollination is less valued may create a tension between the expectations of employees seeking mobility and managers looking to control the talent pipeline.
- Recognize that the new definition of skills is not a unit of measure. While the skills intelligence apps claim to enable workforce planning, this requires bridging from loosely defined skills tags to concrete labor units, a bridge that has proven too difficult to build for most organizations using traditional competency models and will likely have diminishing returns even with the more modern tools as they are today.
Lastly, a more philosophical point about the skills craze. As I grew up in HR, many of us embraced some version of the mantra “hire the person, train the skills”. From today’s vantage point this may sound risky, as “hire the person” sounds like applying undefinable subjective criteria that are laden with our individual biases. However, it is also true that our skills cannot fully describe us or our potential contributions. If we were somehow able to catalogue each and every one of our skills, from tying our shoes to pivoting data or assessing talent or designing the optimal curve of an airplane wing, we would still be missing the personal and professional attributes that form the container for our skills – curiosity, work ethic, compassion, etc. Each of us has our own list of these descriptors, and most organizations have them embedded in their culture, implicitly or explicitly. From my experience as both practitioner and advisor, bringing these cultural dimensions to life can only happen through dialogue – well-facilitated calibration, challenging talent review sessions and thoughtful talent brokering. Through these conversations, enhanced with data, leaders become aligned on a leadership culture that forms the backdrop for productive, skillful work to take place. In designing talent processes around this exciting new category of technology, let’s make sure these aren’t lost.
