VR training data: What it is & how it’s being used

Employee training, like many other facets of the business, has become data focused. With new training approaches like Virtual Reality, companies are tapping into data sets that have gone largely untapped to this point.

Throughout the training process, each person is performing actions that the VR software captures. This includes logging into the system using their name or employee ID, evaluating their virtual environment, answering multiple choice questions, and more.

Discover how Walmart leverages VR data to transform the employee experience.

The cumulative data collected from Virtual Reality provide insight that paper tests and computer-based training methods never have, including where participants focus their attention in a scene and how that translates to engagement and on-the-job performance. Through immersive training data, we’re discovering more about learning than ever before.

Learn more with VR training data

Some of the world’s largest companies are already benefiting from the 5 types of VR training data:

1. Usage data
2. Sentiment analysis
3. Performance data
4. Attention & engagement data
5. Predictive analytics

Read on to learn more about the 5 types of training data that are available when you train employees with Strivr's immersive learning solutions and how all of these metrics can be used to provide much richer training feedback than any prior model.

01. Usage data

Usage data are essential information: who is using the training, how often, and for how long. Ultimately, usage data tell us who completed the training, and it isn’t unique to VR. However, it is especially important for VR training because it helps lead the way toward ROI. Most companies are able to build the strongest business case for VR by implementing modules that affect a vast number of employees, not just a small sample. So the more employees that use the training, the greater the impact.  

Usage data are straightforward quantitative data that set a foundation for all the analytics to come. As you start to analyze training performance, it’s all contextualized by this foundation.

02. Sentiment analysis

Sentiment analysis shows how users felt about the experience in their own words. This is often collected by asking open-ended questions, such as:

• Did you find the training useful?
• Did you enjoy the training?
• Did the training cover the right context?

This is one of the most valuable trainings that I’ve experienced... It makes me feel safe.

Urooj Khalid

General Manager, Verizon Store

We’ve found that VR empowers people, giving them training experiences that they feel actually work. This is powerful. By making your workforce feel like they’re more prepared, you’re also boosting confidence and overall employee engagement.

You want people to experience a situation and get familiar with an environment so they can achieve the same level of performance. In some cases, you want them to feel the same level of pressure that they’ll face in real life. So it’s important to ask these questions to validate the trainees and make sure you’re achieving those goals.

03. Performance data

In general, performance data demonstrate how good trainees are at a given skill of interest, representing a more valid reflection of real-world behavior. However, because of the limitations of traditional data capture (e.g., paper and pencil test, computer-based multiple choice questions), most performance data offer an incomplete, if not incorrect, picture of a given trainee’s competency.

But now with performance data from Immersive Learning, more rich and accurate insights can be derived than from other training methodologies. This is largely due to the unique immersive data captured during training.

04. Attention & engagement data

Attention and engagement data indicate where and how trainees pay attention using metrics such as head movement, eye tracking, interactions, and clicks. These immersive data are unique to virtual environments and offer two advantages compared to traditional training.

• Because immersive training environments closely simulate the dynamic environment of the real world, the behaviors produced by trainees in immersive environments (i.e., trainee choices, how quickly they react, where they look, etc.) serve as a valuable predictor of real-world performance.
• VR collects more meaningful data types more often. Immersive training allows individuals to naturalistically interact with their environments and VR can capture those naturalistic interactions moment-to-moment. This leads to a better understanding of how comfortable and confident – even how expert – someone is with a given task based on the data captured from their naturalistic interactions with their Immersive Learning environment.

Retail training provides a compelling case of how immersive data provides a more insightful story. Most traditional training assesses how good you are at interacting with a customer by having you respond to an imagined customer interaction, oftentimes in multiple choice format. But knowing what to do is very different than knowing how to do it, especially with all of the real-world distractions, emotions, and pressures that are typically present in the workplace. Simply knowing the steps of how to assist a customer is different than actually applying those steps, especially when there may be other distractions.

For example, an associate behind the deli counter has to know a lot of different skills to do their daily job: how to use the deli slicer, how to handle meat, how to do dishes, and more. The really critical piece is how to interact with customers. But with so many operational procedures to worry about, new deli associates can sometimes ignore a customer initially – not on purpose, they’re simply focused on all of those other things. Immersive data tell us why that customer was ignored (e.g., the associate was cleaning dishes) and tee up a coaching opportunity around customer service.

05. Predictive analytics

Predictive analytics are a combination of performance and engagement data mapped to real-world data to create a machine learning based predictive model. While attention and engagement data generally give a better representation of how someone would perform in the real world, it’s with predictive analytics that you can precisely understand how performance in the immersive training will play out once someone is on the job.

For example, you’ve learned through immersive data that a given trainee may be good at spotting safety hazards in a deli, but bad at spotting customers that need attention in general. But what you might not know is precisely how much their poor immersive training performance with customers may impact customer throughput in the real world. Predictive analytics can provide clarity into that relationship, ultimately predicting future customer throughput based on current immersive training performance.

Today, immersive predictive analytics are in their infancy because most companies are not yet measuring data on employees’ real-world performance. But as more companies begin collecting real-world performance data at the employee level, this will allow better and more precise predictive models to be developed. Ultimately, immersive training can be used as a way to assess the workforce and immediately understand the type of real world impact they will have.

It’s important to note that while predictive models can be developed for any sets of data, not all predictive models are equal. The accuracy of the predictive model depends on the strength of the underlying relationship between the data (e.g., the relationship between immersive safety performance and real world safety incidents). In practice, this means if immersive training data is more reflective of real-world behaviors than data from traditional training methods, then the predictive models from immersive training data should be more accurate than those from traditional training data, as well.

Another factor is making sure there is enough data. Again, predictive analytics can hypothetically be built with just about any data sets, but if they are built for data sets that contain too few trainees or aren’t made up of the right types of trainees, then predictive analytics may not apply for data coming from new trainees, hindering the value of predictive analytics.

These considerations and more are what Strivr’s data & analytics team work on for our customers, so that we can properly guide them down the right path to meaningful metrics and ROI. Connecting training to real-world performance is a holy grail for L&D leaders, and with VR, we’ll soon be able to get closer than ever before.

Conclusion

When combined, VR data paint a holistic picture of training at the organization, all the way down to proficiency at the employee level. Equipped with this VR data, you can make better people decisions that ultimately affect large-scale business objectives at the highest level.