Skip to content

This foolproof, 4-step process will eliminate your digital skills gaps.

Gabriel DalportoCEO, Udacity

If you answer “yes” to any or all of these questions, you must read this article:

  • Do you have major gaps in digital skills?
  • Is attrition accelerating and exacerbating your skills gap?
  • Are you finding it hard or impossible to recruit high-demand skills like cloud, AI/ML, and cybersecurity?
  • Is any/all of the above causing business problems—like project delays, lack of innovation, cost overruns, or lost revenue?

Don’t worry. You’re in good company. We work with the largest enterprises across the globe, and they all are experiencing exactly the same challenges you are—companies like Shell, Mazda, BNP Paribas, Airbus, Vodafone, Davivienda, and STC.

The good news is there’s a fast and incredibly cost efficient solution to each of these problems. It also can be an “easy” solution for companies that partner with a “Talent Transformation Platform” (TTP). More on that in a minute.

Whether you partner with a TTP or not, and whether you partner with Udacity or not, the following 4-step process will deliver game-changing results: It’s based on more than a decade of Udacity experience and learnings.

Each of these insights was hard-fought through years of experimentation and refinement. In doing so, we’ve transformed lives, businesses, and nations through massive upskilling and reskilling of workforces. Read on.

There’s hundreds of millions of dollars of value ahead.

Before we start, just what is a “Talent Transformation Platform”?

“Talent Transformation Platform” is Udacity’s way of describing an end-to-end upskilling partnership. It’s radically different from a “learning provider” or an “online learning provider.” Learning providers start and end with some videos or lectures, followed by multiple-choice quizzes. They generally provide good value by being very broad and very inexpensive. Said another way, think of typical online learning platforms as a mile wide and an inch deep—primarily delivering surface-level learning only.

TTPs, on the other hand, work with clients on everything from program design and strategy, to assessments, to customization of content, to customization of learning paths, to cohort-based program management, to linking graduates to projects, to direct ROI measurement.

Now that we have that out of the way, let’s get into the meat of the 4 critical best practices that will solve your digital skills gap and transform the trajectory of your business in the process.

Step 1

Design a customized program for your unique business context.

What doesn’t work: Going for cheap and generic.

What’s the most common mistake nearly every large enterprise makes when the building of a learning initiative for employees kicks off? They don’t start with the mission-critical business problem the program needs to solve.

Rather, enterprises almost always go through multiple phases of rolling out learning programs:

  • Phase 1 is what I’ll call “cheap and broad.” During this phase, employees are demanding access to growth and learning opportunities. In a world of high employee churn and dramatic shortages in the labor market, especially in high-demand digital roles, employers have been put on the defensive and are eager to lean in on learning solutions. However, during this phase, focus is on finding good-quality content that is very broad and very inexpensive. Companies like Udemy, Coursera, and LinkedIn Learning all offer solutions that are viable to this phase. And honestly, every company should have a broad, inexpensive “catalog” of superficial learning topics that can take an employee from 0 to 1.
  • Phase 2 is what I call “online learning disillusionment.” Inevitably, what organizations find when they start looking through their learning data is that:

    • Usage of learning programs is very low. Typically, ~10% of employees will access the tool in a given year.
    • Completion of these programs among people who use them is even lower. Typically, completion rates are ~5-10%, or about 0.5% to 1% of your employees completing a program within a year.
    • Employees who complete these off-the-shelf programs have superficial skills that aren’t applicable to your specific business needs and context. People graduating from a cloud computing program rarely have the skills to be a productive cloud developer. People graduating from machine learning courses are typically wildly underqualified to deploy these skills.
    • Employee churn has actually gone up, not down. That’s because employees haven’t gotten much—if any—practical value from these programs. And they certainly aren’t qualified to take on a new role within their organization. So they get frustrated and leave for a different enterprise that will give them that opportunity.
    • There is literally no measurable impact on ROI. These programs aren’t set up to measure the impact of graduates on the actual business, nor on the business problems they are meant to solve.
  • Phase 3 is what I call “learning to solve the right problem,” which we’ll get into next.

What does work: Building a program around your company’s mission-critical business problems.

Almost universally, companies that start with the big, mission-critical business problem that needs to be solved have transformational employee and business impact from their learning investments.

What types of problems should leaders be solving? These are actual statements we’ve heard from clients:

  • “I have a $2B Microsoft Azure migration I’m driving, and it’s failing because I don’t have qualified Azure developers and I can’t even hire them.” - Top 3 U.S. telecom
  • “We are in an all-out war for cybersecurity talent, and we’re losing. We simply can’t hire them when competing with Google, Amazon, and others.” - Top 5 U.S. bank
  • “We need to deploy machine learning in 70 countries, but have only a small, central data science group that cannot scale, nor understand all the local use cases.” - Top 5 global energy company

As you can see, there are major business problems driven by skills gaps and talent gaps. When you start your learning initiative journey with business statements like these, you know you’re on the right track.

Over time, as your organization achieves success in these initial mandates—by filling skills gaps, reducing attrition, and ultimately delivering on these business challenges—these programs will become second nature, and highly repeatable. And that will enable you to attack the larger problem of continuous learning for a much broader group of employees and internal customers.

What doesn’t work: Making a broad catalog available to all employees and hoping they will magically learn exactly the right skills to fill in your company’s massive skills gaps.

Most companies have a solid understanding of their major digital initiatives and are well aware of what types of positions/skills they will need to fill to execute those programs.

However, you would be surprised and perhaps shocked at how many organizations don’t understand the current state of skills in their existing workforce. Without a talent transformation partner, they have an absence of clarity about what their major skills gaps are. As a result, developing a plan and roadmap for closing those gaps becomes impossible.

So companies go with “Plan B,” which is to make a broad learning catalog available to employees, in the vain hope that they will somehow take the initiative to learn just the right mix of skills required for the company to succeed on its mission-critical business initiatives. Not even once have I seen this strategy succeed.

What does work: Assess the current skills gaps of employees that are preventing you from solving mission-critical business problems.

Let’s take the specific example of the company trying to deploy Azure across its organization. The basic questions you need to ask and answer are:

  • What are the roles/skills I will need to fill to successfully complete this project? Let’s say it’s:
    • 20 cloud architects
    • 25 cloud DevOps engineers
    • 375 cloud developers
    • 15 hybrid cloud engineers
    • 10 site reliability engineers
  • What actual skills do I have within the organization today?
    • If you don’t know, you’re not alone. A large percentage of companies don’t have the data to answer this type of question.
    • Thankfully, the solution is easy. You can either leverage in-house assessment tools, or you can leverage your learning partner’s assessments platform.
  • What people do I have that don’t have the needed skills, but have the basic foundational or prerequisite skills so that they can be upskilled?
    • Again, if you don’t know, make sure your learning partner has a robust assessments platform that can gauge who is most prepared and most likely to succeed.
  • Map the raw material (people ready to be trained) to the open roles, and—voila!—you have your people plan.

BUT: That’s just the start. Now you need to define a custom curriculum that gets employees from A to B.

What doesn’t work: Off-the-shelf programs.

You’re going to learn a lot in your assessments phase. It’s likely you have many employees with a variety of skills and starting points. If you’re trying to get 375 people job-ready as “cloud developers,” you can’t just pick an off-the-shelf learning program called “cloud developer.” That’s because your definition of this role is probably different than everyone else’s. And your implementation is also unique.

So imagine you’ve got 500 potential upskilling candidates:

  • 200 are at total entry level
  • 200 have level 1 skills, but not level 2 or level 3 skills
  • 100 have level 1 and 2 skills already, but not level 3

An off-the-shelf program may take participants from level 1.5 to level 2.5. The net result is that the learning ends up not being right for anyone.

What does work: Tailored and custom programs.

What you need is a program that’s tailored to (1) each cohort’s starting point and (2) your company’s ending point, so you can take each set of employees from exactly where they are to where they need to be. You also may want to throw in some company-specific context, like your particular data architecture, and how that affects the deployment.

You’re going to want to define all of this in concert with your learning provider. Make sure you work with a provider that can meet your employees and company where it is, not where the provider is.

Step 2

Pick a learning platform that delivers outcomes, not videos and lectures.

I can’t tell you how many managers and HR professionals have confided in me that their online learning platforms have been a failure––that people aren’t completing the programs, and even when they do, they don’t have the skills necessary to do the job. And honestly, should this really be a surprise? No, it shouldn’t. Here’s why:

What doesn’t work: Passive learning via videos and lectures.

Ask any world-class software engineer or data scientist: “How did you learn the skills you’re using today?” I will bet you a year’s salary that not one of them will say by reading a book or listening to a lecture. That’s because you can read 100 books on programming in Java or R or Python and not know how to program. You learn to program by writing programs. It’s that simple.

And most (but not all) online learning platforms have totally missed this point. They typically fall into one of two camps:

  • University content brought online: What could be more boring or less impactful than watching a professor drone on in hour-long lectures that you’re going to forget 15 minutes later?
  • Short “how to” snippets: These happen to be super useful if you are an experienced programmer, cybersecurity expert, or data scientist and need to learn how to solve that very specific problem. But these videos are terrible ways to learn how to build a broad set of skills, like becoming a programmer or a cybersecurity expert, or a data scientist. In other words, you’re learning leaves and can never see trees, let alone the forest.

What does work: Active learning via project-based learning.

Learning platforms that start with an employable job resume and work backwards have cracked the problem of delivering outcomes. Specifically, you should have actually done the job or built something meaningful to graduate from the program. A few tangible examples:

  • If you want to become job-ready as an Android app developer, you should have actually built an Android app and had it accepted by the Google Play store.
  • If you want to get a job as a healthcare AI worker, you should have taught AI how to recognize pneumonia off of chest X-rays with better accuracy than a trained doctor.
  • If you want to become job-ready as a self-driving car engineer, you should have had the opportunity to upload your code into a self-driving car and have it drive around a course with your algorithms.

This isn’t science fiction. These are all examples of what’s required to graduate from a Udacity program.

It’s also critically important to differentiate between what are known as “guided projects” and substantive, open-ended projects. The former is a souped-up, “paint by numbers” version of a video lecture. Guided projects tend to be very short and prescriptive. The latter tasks the student with building a substantial app to achieve a desired outcome, without step-by-step guidance. That’s where true learning happens, because the process mimics reality. In the real world, developers solve complex problems, and you simply can’t do that without writing your own code and figuring out your own solutions.

What doesn’t work: Fully online/asynchronous programs and getting stuck.

Speak to any engineer about what it was like learning to code (or to actually work on a project) and they will all tell you that you frequently get “stuck” on a coding problem. When I personally took a full-stack development boot camp a couple years back, this was certainly my experience. And there was nothing more frustrating than working on coding on a Sunday and knowing I couldn’t talk to my instructor until Thursday to help me get “unstuck.” I can’t even imagine the frustration students in 100% online programs feel when they get stuck and can’t move forward.

As a business leader, the last thing you want your employees going through upskilling programs doing is wasting paid company time sitting frustrated, staring at a screen.

What does work: Hybrid synchronous/asynchronous programs with 24/7, on-demand expert help that gets learners unstuck.

At Udacity, we’ve experimented with every conceivable type of program support. And through all that experimentation, the data led us to one extremely simple but compelling answer: The faster you can answer someone’s coding question and get them unstuck, the higher their graduation and success rate. Full stop.

So we set about building a global network of renowned experts who can answer coding questions in under an hour, and even give line-by-line code reviews in about two hours or less—even at 3a.m. on a Sunday morning. No joke.

And that’s one of the main reasons we have among the highest graduation rates in the industry.

Summary of Step 2:

Just remember two things: First, make sure the learning provider you select is built to deliver hard outcomes and job-ready talent. That most likely means the program is built around a lot of project-based work, in which students get to practice their skills on real-world challenges and scenarios. Second, make sure the program has on-demand help from true experts who can get your employees unstuck quickly, so you don’t needlessly waste their time and salary.

Step 3

Support learning with cohort-based program management.

By this stage, you’ve already defined and customized your company upskilling strategy and selected a learning partner who can deliver proven, job-ready skills. Now you’re ready for Step 3—executing the program.

“But wait,” you’re probably thinking. “Don’t I just flick the switch now and let the students do their work?”

Well, you could. But you wouldn’t be very happy with the results. Fortunately, a couple of pretty simple techniques can solve this.

Let me explain:

What doesn’t work: Self-directed learners on their own.

When employees are initially introduced to a learning program, they are usually extremely excited. They see how these new skills will result in their ability to deploy new skills for the company, and that, in turn, often means career advancement for them. So they jump in with gusto.

But these programs take time and effort. Your employees are busy. They have a lot of day-to-day work obligations. They also have a lot of personal obligations (spouse, kids, health, PTA, travel, etc.). All those other obligations are constantly tugging at them. It’s very easy to deprioritize their long-term success (the results of learning a new skill) to satisfy all those short-term demands.

That’s why just providing a learning solution to an employee and leaving them on their own often results in very low completion rates and very high frustration rates.

What does work: Cohort-based learning.

When you onboard an employee to a learning program with a group of other employees at the exact same time, and you connect the employees together, and you track their progress together, and you set a shared completion deadline, several behavioral psychology factors kick in, to your favor:

  • The first is pretty obvious: If you’re going through this with your peers, you feel like you’re part of a community. You have a group of people you can rely on for support. And you’re energized to make progress along with them.
  • The second is competition. When you see others moving ahead, we as humans have a natural competitive drive to catch up to the leaders.
  • The third is peer pressure. You don’t want to be the only one who falls behind or who doesn’t complete the program.

All these psychological factors are intrinsic to a cohort-based program with strong community features and moderation. The takeaway? Make sure your learning platform makes cohort-based learning, enrollment, tracking, and administration automated and easy.

What doesn’t work: Full-time work and reskilling, simultaneously.

It is unreasonable to pretend employees can make major upskilling investments and work 40 hours a week. We’ve seen it time and again in the data. Everyone loses: Graduation rates drop. Employees get burned out. And the risk of churn is elevated.

And without manager buy-in to the program, everyone is going to be frustrated.

What does work: Management support with protected time.

One of our critical best practices is manager support, and especially with protected time. The first part of this is getting the employee and manager on the same page with a supportive relationship and built-in accountability. When the manager has access to a dashboard and can track employee progress, they can have meaningful conversations around blockers (either technical, or simply time available) and how to stay on track.

The second part is creating dedicated/protected time for employees. This is crucial and often ignored. When I have a client come to me and talk about low graduation rates, I always ask three questions, and one of those questions is “Are you providing protected time?”

If you really want to accelerate your employee learning and application to your business, try embedding learning along with day-to-day work. For example, while an employee is learning about data visualization, work with a manager to find opportunities to apply those data visualization skills in real time. That has the added benefit of providing instant ROI, as well as immediate reinforcement of learning.

Step 4

Deploy graduates on real business problems.

The final set of best practices may seem a little less intuitive, but are no less important—especially if you care about company ROI and employee retention.

What doesn’t work: Doing nothing after graduation.

One of the biggest mistakes a company can make, often with terrible consequences, is to launch a digital upskilling program and not tie it to opportunities to use those skills immediately after graduation.

Put yourself in a learner’s shoes: You’ve just invested multiple months into learning a state-of-the-art new skill. Let’s say it’s machine learning. And let’s say you were still working your normal job, and even with some protected time, this effort wound up eating into your personal and family time. So it was exhausting, but you made it to the promised land. What is the one thing you want to do more than anything at this point? That’s right—you want to go build a machine learning model in the real world.

Now, imagine you go to your boss, and she doesn’t have any relevant ML project for you to work on. You look around the company and apply for a ML job, but your employer doesn’t support the transition.

What happens next? That employee leaves, and the company doesn’t achieve a measurable ROI on the reskilling investment. In fact, it may even end up being negative ROI.

What does work: Tying new skills to new job opportunities.

Now imagine the opposite situation: You complete your machine learning skilling, and your boss has a great project that they think a fresh ML analysis could break open. Or better yet, your boss promotes you into that open ML position, and you get to stretch and grow and use all these new skills on interesting problems.

What’s the outcome for you as an employee (and for the company)?

  • You get to apply your newfound knowledge to a practical business problem, which is exciting and rewarding for the individual. But as importantly for the company, the cumulative impact of many employees deploying these skills immediately results in an explosion of measurable ROI. Time and time and time again, we see that the strongest ROI from reskilling comes from graduates directly applying that new skill to a novel problem at their company, resulting in step-change improvements in cost, productivity, or revenue. The average directly measured ROI for Udacity clients? 600%.
  • You are growing in your career, tackling new challenges. You’re now thankful and emotionally indebted to your employer and manager. And you are much more likely to stay with the company. In fact, based on our extensive data, at this stage, clients see a reduction in employee attrition of 5-10 percentage points (not to be confused with percent).

What doesn’t work: Measuring success with meaningless metrics like “learner engagement” or CSAT.

This is a big paradigm shift for most companies. But success should not be measured by learner engagement or CSAT. These measures, while interesting, are totally uncorrelated with our primary goal—closing your massive skills gaps so you can achieve mission-critical business goals. I’ll trade 1,000 high CSAT scores for 2 learners who applied their skills to a big problem and saved the company $100M (a true story).

What does work: Direct ROI measure.

This is absolutely critical. Measure direct ROI from the program. There is an “easy route,” which is better than nothing. But there is a much more powerful, “harder” measurement that will shake the foundations of how you think about investing in talent.

The “easy” measurement: Directly measure how many employees were successfully deployed into unfilled roles/projects because of prior skills gaps. This will result in direct ROI from:

  • reduced recruiting costs
  • reduced employee churn

The “harder” measurement: Track what your learners deployed their new skills on post-graduation, and work with them to size the business impact (more revenue, lower costs, etc.). This may sound hard, but it’s really not. It just requires a platform that automates the information-gathering and aggregates the insights into a total ROI.

Make sure to pick a partner that can both help you think through your deployment strategy as well as provide automated tools to measure and track the ROI. At Udacity, for example, we’ve built tools that track learners after graduation and directly measure the impact of deploying their new skills on a novel project.

The promised land and how to make this easy.

We’ve deployed this approach across hundreds of Fortune 500 and Global 2000 clients. It is not a stretch to guarantee that anyone who follows this playbook is going to achieve:

  • typical graduation rates between 60-90%
  • a massive reduction in employee churn/attrition by 5-10 percentage points (i.e., 30-50% reduction in attrition rates)
  • an explosion of skills deployed on mission-critical projects, resulting in: directly measurable acceleration of innovation, increases in revenue, and reductions in cost
  • an average, directly measured ROI of over 600%

The real question you’re probably asking yourself is: “This seems hard. Can I pull this off?” The answer is yes. But it’s a heck of a lot easier if you have an enabled partner who can guide you through each step of the process, and leverage a world-class technology platform with best-of-breed learning programs. Udacity has spent over a decade developing technology, tools, and people custom-made to deliver on each of these best practices.

Find out more about how talent transformation can eliminate your digital skills gaps today.