Is there a wrong or right way to learn tech skills such as coding or Big Data analytics? According to a Harvard Business Review report, only one in four senior managers believe that training is essential to business outcomes; although U.S. companies spend about $160 billion teaching new skills, people aren’t absorbing knowledge without hands-on instruction. As Fast Company reported, book learning or memorization alone may not be the most efficient ways to remember key concepts. But even if book learning can’t be relied on without actual practice, it’s still important, especially when it comes to learning data science, said Cushing Anderson, vice president of IT education and certification research at IDC: “If you're going to become a data scientist, there are probably books you need to read, but that's insufficient to becoming a data scientist.” So what’s the right way to learn tech skills? Here are some winning strategies from experts.

Shadow a Mentor

One of the most effective ways to learn is to observe a mentor in a real setting, especially if it involves actual tasks. For example, you could shadow a data scientist running an analysis, or a sysadmin performing regular maintenance on a network. “Go stand next to the person who's doing that job a little bit, help them out, be their apprentice, so to speak, so that you get to see the kinds of things they're doing,” Anderson said. “We call that informal learning, and it works. It's very functional.” When you go to a class after that practical training, he added, you’ll have a framework to understand the concepts.

Break Down Skills Into Microbehaviors

When learning a new tech skill, start by breaking small concepts into “microbehaviors,” said Martin Lanik, CEO of leadership development platform provider Pinsight and author of “The Leader Habit: Master the Skills You Need to Lead In Just Minutes a Day.” “Because those are easier to practice, they take less time and it's easier for the brain to model that behavior and to turn it into habit so they become as automatic as brushing your teeth, taking a shower or making coffee in the morning,” Lanik said. “That's the basic idea behind microbehaviors.”

Train for Programs You’re Passionate About

Whether the skills you need to learn are for cybersecurity, data science or cloud engineering, the classes you take must be relevant to what you need to learn. If you take a class for a tech skill you can’t relate to, you won’t retain the knowledge, noted Jim Johnson, a senior vice president at IT staffing firm Robert Half Technology. “My biggest fear is that I leave and come back in three months and I've got to redo it again because I didn't get any of it to stick,” said Johnson, who has been providing tech training for recruiting and sales positions. “People are going to engage with what they identify with, and if they don't identify with it today, they lose it.”

Be Flexible with Your Training Methods

Tech pros should try a variety of methods to stay sharp, whether it’s online learning, videos, books or boot camps. The key to successful learning is to be “willing to be flexible and not engaging in a one-size-fits-all approach to your training—both as a user as well as a manager,” Johnson advised. Vendors such as IBM offer skills training programs that combine online and face-to-face learning along with actual project work. In its first three months, the IBM Blue Collar Program covers basic math, how to code Python, and the basics of machine learning, noted Seth Dobrin, vice president and CTO for IBM Analytics. “By the end of that three months, and actually during that three months, they start working on an internal IBM data science project,” Dobrin said. “While they're learning to code in Python and learning about machine learning, they actually are also starting to execute a project that's important for IBM's internal efforts. They're putting to practice what they're learning in real time, and it's not just a book-learning exercise.”

Stay Current by Attending Conferences and User Groups

Tech pros looking to learn new skills and pick up on new trends should attend conferences and user groups. “You may not learn to be a JavaScript developer in a 30-minute presentation, but at least it gets your mind thinking about things and allows you to come back and make recommendations to your manager or potentially your staff based on what we're seeing as market trends or opportunities,” Johnson said. At IBM, Dobrin requires his team of data scientists to stay current on the latest literature as well as the newest techniques and software languages. And as a technical leader, he also participates. “That's really important to still get your hands as dirty as you have time for so that you really understand what's going on, because you quickly become obsolete,” Dobrin said.

Apply New Skills Quickly

When you learn new tech skills, don’t just file and forget them; you’ll need to practice what you learned. Dobrin refers to this process as converting new learnings into “muscle memory.” If you attend classes on learning how to code and writing a machine-learning pipeline, “understanding them conceptually is completely different from actually applying them,” Dobrin said. People often have the wrong perception of what training is designed to achieve, according to Johnson. “It's not the training that’s bad,” he said. “Training has a marketing problem. People come in with the wrong expectations, and that's essential to know.” Is there really a wrong way to learn? Not necessarily, according to Johnson. “The only wrong way to learn is to not learn,” he said. “You have to be willing to evolve your skills, look at different ways of doing things and kind of embrace the constant learning as part of what it takes to be in IT.”