A version of this article originally appeared in Chief Learning Officer.
We live in a world where movie-grade cameras fit in pockets and USB thumb drives can hold a lifetime of content. The commoditization of professional video equipment and storage has enabled L&D teams to create exponentially more video content and scale their training programs at a fraction of the cost it would’ve taken just a few years ago.
While capturing video can help your team conduct trainings more efficiently, how can you ensure that their training videos provide the greatest value for the rest of your workforce?
The video content management system provides an answer. In recent years, video CMSs have begun to incorporate machine learning technology, enabling the software to learn by identifying patterns in large data sets. Specifically, machine learning is enabling organizations to unlock the value of video by helping employees find the information they need to perform at their best.
In traditional learning management systems and other content repositories, video search is typically limited to manually entered metadata like titles, tags, and descriptions.
For your employees, this presents two problems. First, it limits the chances that their searches will actually find the right video. Second, it prevents employees from searching within videos to find that relevant two-minute segment they need to accomplish a task.
Consider a 45-minute-long recorded instructor-led training session. On average, a trainer speaks 125 words per minute, so during the session, let’s assume more than 5,000 words will be spoken. Of those, let’s conservatively estimate that only 10 percent will be of unique value to an employee searching for information.
That’s 500 words. Manually tagging them all would take hours. Not tagging them would limit the discoverability of the video. Neither outcome is desirable.
Even if the trainer painstakingly added 500 tags to the video, those tags would still only help employees find the starting point of the recording. In most cases, your people don’t want to watch training videos in their entirety. Instead, they’re looking for precise moments in the video that contain insights on a particular topic. With traditional video search, the only solution is to click randomly through the timeline or take the time to watch the full recording.
These inefficiencies in search reduce the value of your e-learning initiatives because they prevent employees from using video as a just-in-time learning resource.
Fortunately, video search has made tremendous advances in the past few years, and these advances have become standard in many video CMSs. Modern video search engines enable employees to find any word spoken or shown on-screen within any video and then instantly fast-forward to that precise moment.
Under the hood, these video search engines are powered by machine learning. Specifically, two technologies analyze video content to create a search index: automatic speech recognition works to recognize words spoken by your instructors, and optical character recognition discerns the words presented onscreen. These technologies are trained with massive data sets and deep-learning algorithms to identify individual words and phrases, which are then time-stamped and added to the search index.
The result? By converting speech and on-screen content into a timeline of text, these search engines make video content as discoverable as documents or email.
When your employees search your video library, they not only find the most relevant recordings, but also the precise moments within those videos where the relevant topic was mentioned by the instructor or shown on their screen.
By making your training videos searchable, you unlock all the valuable information they contain and make it available to your employees at their moments of need. This can dramatically improve workforce productivity by reducing the time employees spend searching for information required to do their jobs.
At companies such as Synaptics, the results have been measurable. The organization estimates that employees save 15 minutes a week through the use of video search. That modest individual gain translates into thousands of hours of productivity gains across the business each year.
As your organization continues to create e-learning videos, remember that the content is of little value if it isn’t discoverable. Machine learning technology provides a novel way to realize the latent value of these recordings. Learning leaders who capitalize on this opportunity stand to improve workforce productivity, drive business results and ultimately improve the bottom line.