The Ultimate

Skill Data
Handbook
 If you’re like most HR, talent, and learning professionals, you’re
looking for new ways to be a business enabler, which means
being more data-driven and strategic about people development.
Organizations that prioritize data-driven decision-making are
twice as likely to have significantly exceeded business goals.

Table of Contents Companies too often view corporate learning, talent, and HR
technologies as an expense rather than an investment in growth.
But that’s an outdated narrative, and it’s changing fast.
What is skill data? 4
On top of their traditional responsibilities, people teams are now also
Where does skill data come from? 6 expected to enable revenue growth and business agility throughout
the employee experience. They’re expected to:
Good vs. bad skill data 8
Where should skill data be stored? 11
Provide career
Reallocate Generate
How can skill data be used? 13 growth and
guidance for people quickly definitive ROI on
individuals, when business investments in
Differentiating skill data technology 15 teams, and
managers
priorities shift people

Skill data glossary 17

Identify skill
Develop a Maintain an Decrease
gaps in the
more agile internal talent spending on
workforce and
workforce marketplace external hiring
close them

Support a Decrease stress

Create an Reduce bias in
more diverse and improve the
active learning career
and unbiased mental health of
culture opportunities
hiring process employees
 Managers are under more pressure than ever. Constant business
disruptions brought on by the pandemic have executives looking to
L&D and HR teams to keep their workforces agile and ahead of the
competition — and to do more with less.

So how can you prepare your workforce for the future?

It starts with skill data.

Skill data can provide visibility into the supply and demand for
your organization’s skills. And it can help you align your workforce
with your business imperatives, save money, and increase
employee engagement.

But it’s not simply about accessing data; it’s about understanding
your data sets to use them correctly, responsibly, and efficiently.
So you can bridge the gap between learning and talent initiatives
and organizational goals.

The field is still emerging — and evolving fast. We’d like to provide
some clarity to help L&D and HR leaders make better informed
Right now, only 37% of companies consider
investments with skill data. However, we know there are not yet
common standards around this topic, so this guide also aims to themselves to be highly insight-driven, according to
provoke debate to advance the state of the art, which is not yet a recent study. Understanding and communicating
meeting the demands of 21st-century businesses. For those with
your skill data needs will be the key to staying ahead
something to add to the subject of skill data, we encourage
feedback and welcome discussion. of disruption and your competitors.

The Ultimate Skill Data Handbook 3

 1. What is skill data?

Skill data is the measurement of what

your people can do, ideally aligned
with the work that your organization
needs to get done.
We like to think of skill data like medical or health data. There
are endless measures relevant to your physical health, from your
smoking history to your injuries, cholesterol, and exercise habits.
Depending on your goals, you can focus on tracking and improving
different areas of your health. For example, losing weight and
bodybuilding are two very different goals that will require different Skill signal:
strategies and metrics. An indicator that a person has a skill, and in many cases,
the skill level. Signals can include skill ratings, content
Skill data can help assess the general health of your organization. It
consumption, results from third-party assessment tools,
can help uncover your areas of weakness and point you towards the
previous or current job descriptions, and more. These signals
right steps to improve. Like health data, skill data can live in many
different places, and is represented in all different ratings, scales, can help technology and administrators make inferences
and variations. To put skill data to work assessing your organization about all the skills you have and those you’re building.
holistically, you need to connect a variety of different sources.

The Ultimate Skill Data Handbook 4

 The most obvious, direct examples Content data: What your people are
consuming can tell you what areas they’re
Activity data: This is mined from workers’
daily behaviors, including platform logins, web
of skill data are skill assessments, focusing on. Are high performers all using searches, and content views. This data can
similar content? What does the utilized help administrators make inferences about
skill ratings, and skill endorsements. content tell you about the skills being built? emerging skills that can help their business.
Maybe some content providers are extremely Perhaps technical teams are learning a new
However, other metrics can also be popular, and others are rarely used. Content programming language or the sales team is

useful to identify and quantify skills. data can tell you where to invest further or
even which providers to cancel.
using a new tool.

Degreed calls these skill signals and User data: This is data about the users — your
Assessment data: Many systems can offer
skill assessments, but the most intelligent
they can include: people — and their work experience. It can be systems will be able to help you identify which
taken from learning profiles, resumes, third skills are changing. You’ll know if a key skill
party networking sites, and applicant tracking is improving across your entire company or
systems (ATSs). Providing a retroactive view, if a cluster of other skills is increasing within
it can offer visibility into people’s past projects individual teams.
and hidden skill sets that wouldn’t ordinarily
surface in their daily work. Maybe one of Much of this information you may currently have
your project managers has a graphic design access to — because your people are already
background that she doesn’t use in her current generating it. But to unlock the real power of this
role but could make her an excellent fit for an data, you’ll need to collect, organize, analyze,
upcoming campaign that needs help. User data and use it.
can tell you that.

The Ultimate Skill Data Handbook 5

 2. Where does skill data come from?

Here’s the short answer: everywhere. Let’s go back to the health example. People are
generating all kinds of health data every hour of
daily steps. Your browser search history knows
what recipes you’re downloading. Your financial
every day. The challenge is aggregating all of that statements show what food you buy. Dieting apps
The long answer: Anywhere your data and making sense of it so individuals can
view their health profiles holistically, identify
capture information about your daily nutritional
intake, and personal trainers know how you’re
people interact or input information warning signs, and reach their goals. Skill data training for that upcoming marathon. That’s a lot
has the same challenge. of data, in a lot of different systems.
can be a source of data on their
Many vendors and organizations start with skill If the HCM system is like a doctor, all the other
abilities and experiences. This is data from a human capital management system systems your people interact with are the learning

both good and bad. (HCM). At Degreed, we think of HCMs like doctors.
You might visit your doctor several times each
version of smart watches or workout coaches. Find
some examples of these systems aligned to the
year but certainly not every day. And while doctors employee lifecycle on the next page.
measure your height, weight, family history, vital
signs, and acute health issues, they don’t know No single system can supply all the necessary,
the specifics of your daily diet, sleep, or exercise accurate, and unbiased data you need to make
routine. What your doctor sees is important, but it’s decisions about your workforce. And in the age of
not the entire story. APIs, you don’t need to limit your company to one
system. The most agile companies use all of
Now think of all the data created between your these platforms together by synchronizing skill
doctor visits. A smart watch can tell you about your data for analysis.
sleep patterns, calories burned, heart rate, and

The Ultimate Skill Data Handbook 6

 Skill Data Collected Throughout the Employee Lifecycle

Performance New Growth

ATS HCM Reviews Content Platforms Opportunities
Applicants submit resumes. HCM systems collect skill Weaknesses, achievements, Platforms such as Coursera or Workforce technology can identify
But only 21% of workers say data from new hires. But guidance, and qualitative Pluralsight show course opportunities for workers to use
their profiles are up-to-date only 2025% of workers feedback. But 61% of completions, searches, and their new skills or their old skills in
with their skills. fill out their company’s employees don’t get regular emerging skills. new capacities. Internal gigs,
talent profile. feedback on their skills. mentorships, stretch assignments,
or full time roles.

Onboard Engage Upskill Develop Reskill Accelerate Extend

Behavioral The Employee Lifecycle

Assessments LXP LMS Assessment Tools
Soft skill assessment,
Systems That Capture
Your people’s focus skills, Learning data, content This can include credentials,
learning personas, etc. skill ratings, social sharing, people are finding and badges, skill assessments, and
and Use Skill Data
and mentorships. sharing, compliance, etc. they can help establish a baseline
of your org's skills. But 40% of Source: Degreed, The State of Skills 2021
workers do tasks outside their
main job descriptions, and likely
won't rate those skills.

The Ultimate Skill Data Handbook 7

 3. Good vs. bad skill data
Just like other data sets, not all skill data can be trusted equally. Validity
Your data’s validity is contingent on your sources and the methods
used to generate it. Is your information directly verified, reviewed,
While good skill data can help you make smart decisions, bad or certified? Perhaps it’s inferred or pulled in from other systems.
skill data can lead you to inaccurate assumptions. Reliable data Are those systems trustworthy? By asking these questions, you can
begin to account for bias and trust. For each source, there are a few
starts with validity, but it can also require other attributes, like nuances to consider.
volume and freshness. Ratings: Ratings offer a lot of data very quickly, making them
an ideal starting point. They take only seconds to complete;
your people simply rate themselves numerically on a scale
for each skill. Though as a self-report measure, they can
include some bias.

Assessments: Assessments bring a level of objectivity because

they’re rooted in demonstrable ability. They’re more reliable but
take longer to complete, like a test. With assessments, think
about the tools you’re using and how they speak to each other
(or don’t). Different types of ratings with different terminology
will only muddy your data.

Inferences: We can now go beyond rated or tested ability to all

the data points within a professional career. Inferences are taken
from skill signals and offer more insight into peripheral skills that
workers might not choose to assess or rate. A new email marketer
Bad data Good data just joined the team and only chose to assess skills pertaining to
her role. She didn’t mention her Salesforce skills, but these can
be discovered from inferences.

The Ultimate Skill Data Handbook 8

 Volume
Sparse data can be misleading. For example, if your workforce has 1,900 people and
you see 160 data points that show excellence in a particular area, you’re likely to
assume your workforce is strong in that field. But the remaining 1,740 people may have
very low proficiency in that area, and your organization could be missing a crucial
growth opportunity.

Diversity: The more metrics you track, the more comprehensive your view of your
organization will be. With a wide-angle lens, you can identify patterns and create
early warning systems that alert you to disruption. In other words, you can detect
small ripples before they become disruptive waves.

Freshness
Is your data updating as quickly as your people’s skills are? Their abilities can
change by the month, week, or even day. Fresh data comes from constant workforce
engagement. So if you’re not capturing data from systems your people use regularly,
your data may be inaccurate or out of date.

The Ultimate Skill Data Handbook 9

At Degreed, we incorporated validity, Build Your Profile
volume, and freshness into our
upskilling platform design. We collect
diverse information from a range of Skills  Add
39 Skills Added
sources to help build a real time view Top Focus

of workforce capabilities. 28 Skill Ratings

September
4 Focus Skills

133 4 Experiences
Skill Signals

3 Achievements

32 Learning Items

Supporting Info Edit

Community Relations 7 7 7 +4 38 Signals

Resume

Email Blasts 7 6 6 +2 32 Signals

LinkedIn

Sales Database 6 6 5 24 Signals Personal Website

The Ultimate Skill Data Handbook 10

 4. Where should skill data be stored?

Traditionally, organizations have used two main compliance training. That means that employees And for that reason, a single system of record for
sources as the system of record for skills: learning are likely interacting with HRIS and HCM systems skills — one definitive system for all your skill data
management systems (LMS) and human resource very infrequently — like doctor visits — and that — wouldn’t be useful. That’s why it doesn’t exist.
information systems (HRIS) or their cousins, limited engagement means limited data.
human capital management systems (HCM). Instead, innovative organizations are beginning to
For a comprehensive view of dynamic skill use more open systems that integrate (or better
The LMS, of course, has long been used as the data, this question of storage becomes more yet, translate and synchronize) skill data from
system of record for learning. With access to complicated. The answer requires constant user various systems. And in doing so, they’re distilling
employee names, course titles, completion dates, engagement, a diversity of data, and actionable that data into a structured, common language
assessments, and certifications, an LMS can store insights. Right now, there is no universal system that’s readily accessed for analysis, even in a
reliable reports. And many include a basic, static of record for skills. Skill data can look entirely different talent or business analytics tool. This
list (or taxonomy) of skills that can be used to different between and across systems, vendors, creates a living, evolving map of the relationships
tag content and people. However, these lists are and partners; they all have their own labels and between different representations of skills across
limited in how they can impact the business. definitions for skills. platforms and systems.

HRIS and HCM systems are designed to manage Imagine trying to integrate and manage all that
critical administrative HR functions like payroll, disparate data in a single system. If your system
benefits administration, time and attendance — of record even allows it (few do), it would be
and in some cases, performance management and duplicative, messy, inconsistent, and noisy.

The Ultimate Skill Data Handbook 11

 Learning systems Business learning content Public learning content
What matters most, then, is not a Learning Program Virtual Custom
Course Video Business Online
Management Management Training Learning Podcasts

single centralized system, but an

Libraries Libraries Media Videos
Systems Platforms Systems Programs

intelligent, integrated ecosystem Video CMS

Collaboration Learning
MODCs
Professional Certification Research
Blogs Books

that can access as well as enrich

Tools Analytics Journals Programs Reports

any of the different systems

where skill data is stored and
used. Let’s look at how Degreed HCM Skills Digital Skill

is able to do that: Data on Systems Taxonomies Badges Assessments

Data on
demand supply
for skills Career
Pathing
Labor Market
Data
Feedback +
Ratings
Talent
Intelligence
for skills
Tools

Coaching + Talent Applicant Career Vendor

Mentoring Management Tracking Development Management
Tools Systems Systems Tools Systems

Opportunities to put skills to work

The Ultimate Skill Data Handbook 12

 5. How can skill data be used?
You can use skill data to develop future skills, find
growth opportunities, identify organizational
strengths and risks, and more.

To begin using skill data, start first with your

personal, team, or organizational goals. Think
again about physical health goals: Are you trying
to lose weight or gain muscle? Run a marathon or
rehabilitate your knee? Lower your cholesterol
or increase your endurance? All of these goals
will require different (sometimes opposing)
metrics and strategies.

When you identify your goals, you can begin

finding and tracking the metrics that matter
to you. And skill data is not just valuable to HR
teams; it can be useful to individuals, managers,
and business leaders, too.

The Ultimate Skill Data Handbook 13

HR teams Individuals Managers Business leaders
In a recent Gartner survey of learning and Individuals won’t use tools voluntarily unless they Managers are one of the most critical links Business and function leaders are usually aware of
development leaders, 71% said that more than gain something in return (think Instagram/social connecting the workforce to business priorities. revenue goals and organizational transformation
40% of their workforce has needed to develop new validation, Netflix/entertainment, or workout apps/ And yet Degreed and Harvard Business Publishing efforts. But far fewer know how those goals
skills since the start of COVID-19. These types feedback). And individual engagement is the key research found that 61% of workers want more translate to the skills your organization has or
of workforce disruptions will continue to cause to gathering data. So what do individuals get from guidance on aligning learning opportunities with those it needs. With visibility into the supply
massive shifts in required skills. HR teams can stay using, generating, and accessing their skill data? their skill gaps, and 45% of them look to their and demand for skills concerning the work your
on top of disruption — and be ready to seize new They can: managers for that guidance. So with skill data, business needs to get done, business leaders can:
business opportunities — by monitoring early signs managers can:
Identify the skills they want or need to develop Build innovative and creative strategic
of changes in skill data. As an HR leader, you can:
in the future Identify high performers and those with workforce plans aligned to business
Identify areas with changing priorities and potential opportunities
therefore shifting skill sets Demonstrate their growth
Connect teammates who have similar or Identify the best avenues for accomplishing
Distill roles and projects into skills and Discover new and relevant content complementary skill sets projects between employees, contractors, gig
outcomes to increase agility and collaboration workers, partners and vendors, volunteers,
Find mentors or subject matter experts in areas Help their direct reports identify growth robots, and software
Provide personalized upskilling based on they want to grow opportunities, increasing engagement
individual interests and goals Increase organizational agility and adaptability
Find internal opportunities like new projects, Assign projects to appropriate team members through workforce redeployment plans
Create an internal talent marketplace to enable internal gigs, or even full-time roles based on skill profiles
career mobility Stay competitive by cross-referencing labor
Find the right internal (or external) candidates market data to spot emerging trends
for open roles on their teams

The Ultimate Skill Data Handbook 14

 6. Differentiating skill technology
There’s a lot at stake when selecting upskilling technology for your organization.
The three primary functions of a dynamic skills management platform are
generating, managing, and analyzing your skill data. Some platforms do these
well — but others do them even better.

Generate data Analyze data Manage data

Understand the skills we have and the ones Inform decisions and enable action based on skills. Connect skills to work, people, content, and
we need. other skills.
Good - Creating insights into the supply of
Good - Using AI to infer skills based on existing skills to inform HR decisions Good - Organizing and connecting skill
internal and external data sets data within a single HR system
Better - Exposing demand for skills and
Better - Keeping skill data up-to-date enabling individuals, managers, and Better - Synchronizing and exchanging
through user engagement and integrations business leaders to take action skill data across all relevant HR and
business systems

7
5
3

The Ultimate Skill Data Handbook 15

 What next?
Choosing and implementing the right upskilling technology for your
organization will be the key to effectively using your workforce skill
data. But according to Deloitte, creating a data-driven culture is even
more important than buying the right tools or finding the best talent.
Jane Stevenson
3 Skills
Right now, only 37% of companies rated themselves highly on
“insight driven maturity”, according to a recent study. Understanding Skill Focus Skill Self Rating Manager Rating
and communicating your skill data needs will be the key to staying
ahead of disruption and your competitors. Branding Strategy
4 4

Learn more about how Degreed can help you understand, build, 5
Invision

and activate the skills your workforce needs next.

6
UI Design

Visit us at explore.degreed.com or contact your Degreed

representative today.

The Ultimate Skill Data Handbook 16

 Skill Data Glossary
Skill Data Basics
API Data science Machine learning
Definition: An intermediary between two or more Definition: A variety of scientific methods Definition: Machine learning is a subset of
systems to communicate with one another. used to find insights from large amounts of artificial intelligence. It can leverage data to
structured and unstructured data. It’s dedicated identify patterns and make decisions with little
Why it matters: An open upskilling platform to collecting, storing, and analyzing information or no human intervention.
that works with APIs can share and consume about people, machines, and the wider world.
skills, skill ratings, and skill data from any other Why it matters: An upskilling solution that uses
platform that also has the necessary APIs. Why it matters: Data science is about enabling machine learning will be able to provide better
Skill data from APIs is typically used to keep companies to make key strategic decisions content recommendations but also help identify
a user’s profile in sync with other enterprise based on informed analysis. It defines and trains emerging skill sets your workforce needs, find
applications or communicate your skills across skill models which upskilling platforms use to skill gaps, and indicate weaknesses to focus on.
your organization. personalize content and experiences for users.
Skill Data
Skill Model Artificial intelligence Definition: A data point that relates to the
Definition: A process that takes a set of inputs Definition: Software sophisticated enough capability, demonstration, or definition of a skill.
(data) and makes predictions about a user’s to replicate the abilities of humans. AI is how In short, skill data is the measurement of what
future behavior based on a historical dataset machines learn from experience, adjust to new your people can do.
of similar users. inputs, and complete tasks.
Why it matters: Skill data has many different
Why it matters: Sophisticated skill models — Why it matters: Sophisticated systems can facets that can be used to make informed
utilizing data algorithms — can identify relevant use skill models and data science to continue business decisions. It can include a skill, its
and personalized content, recommend new personalizing the learning experience through definition, assessments measuring the skill,
skills to learn, find subject matter experts to artificial intelligence. It can do things such as ratings of the skill, or the relationships between
follow, and more. These models will be most auto-populate learning plans or suggest internal these data points. This can inform business,
helpful when they have access to large amounts candidates for open positions or projects. hiring, performance, and talent strategies and
of data. This means that the more engaging enable more precisely targeted investments
your upskilling platform is, the more data it will in recruiting, workforce planning, capacity
acquire and the better it will perform. management, and change management.

The Ultimate Skill Data Handbook 17

 Organizing Skills
Skill Taxonomy Skill Graph Skills I/O
Definition: A hierarchical system of classification Definition: A visual representation of a skills Definition: Degreed’s Skills I/O can manage
that can categorize and organize skills. Unlike ontology, a skill graph highlights skills, their skills during their lifecycle. You can use the
a skills graph or skills ontology, a taxonomy is relationships, and distance between one Skills I/O to build taxonomies, manage multiple
more structured and almost linear in its format another. skills sources, and your organization’s skills.
and typically includes definitions.
Why it matters: Understanding how different Why it matters: Whereas taxonomies, ontologies,
Why it matters: This can help people understand skills are related to one another can inform and graphs help us understand and organize
the skills they have, how they relate to how artificial intelligence and models offer skills, our Skills I/O can put those concepts
organizational needs, and what they should upskilling opportunities. into practice.
learn next. It can also show what skills are
included in other skills. Skill Cloud/Skill Inventory/Skill Registry
Definition: An inventory of skills across
Skill Ontology organizations that includes all known skill terms.
Definition: A set of skills and relations. It’s the data set that’s used to evaluate skills to
have in organizational skill lists, ontologies,
Why it matters: A skill ontology allows or taxonomies.
organizations to define and measure
relationships between skills. It helps create Why it matters: A skill cloud helps organize and
a common language and understanding of standardize skills across an organization, but
skills across a variety of different dimensions. a skill cloud alone does not make these skills
For example, across individuals, teams, and actionable.
companies, the definition and terminology
used to describe a UX designer will vary. A skill
ontology is able to aggregate all of that data
and recognize that different systems are
talking about the same entities and build
relationships between them.

The Ultimate Skill Data Handbook 18

 Managing Skill Data
Skill Analytics Reporting
Definition: Deriving meaning from collected skill Definition: The process of gathering and
data. presenting an accurate analysis of the data
collected.
Why it matters: Analytics can help you identify
patterns, trends, strengths, weaknesses, and Why it matters: Gathering, analyzing, and
other significant indicators to raise awareness pulling insights from data will not make it
about the larger organization and your people. actionable. Learning to report and present your
findings in accessible ways will allow you to
Insights communicate the importance of the changes
Definition: Inference or prediction of what comes you wish to establish.
next based on the analysis of data.
Skill Data integration
Why it matters: Data itself cannot tell you how to
Definition: The communication of tools that
make improvements to your existing processes.
produce or store skill data, including HCM
It’s the way you interpret your data, apply it
systems and skill assessment tools.
to your business, and allow those patterns
and indicators to preempt new challenges, Why it matters: Skill data integration will allow
opportunities, and needs in your workforce. for a more comprehensive and accurate view
of individual and organizational skill levels by
pulling skill data from systems in the flow of
work. It will also enable “skill signals” for a user,
offering a richer picture of their skill level.

The Ultimate Skill Data Handbook 19

degreed.com