Use Cases For xAPI

Use Cases For xAPI
Lee Charlie /
Summary: The xAPI offers a variety of uses. This guide will analyze and explain three of the most common scenarios.

How Is xAPI Used?

There are many ways to use xAPI. xAPI is a document that, if you follow it, will give you a method for how you can communicate learning activity data between one system and another. It is a derivative of the term ‘API’, or Application Programming Interface. Before the advent of API’s, getting two systems to talk with each other required a large amount of code development and integration time. API’s have transformed this relationship, meaning that basic integrations between different software packages are now relatively trivial.

eBook Release:
eBook Release
Download this eBook and discover why and how you too should be using xAPI.

There are many use cases for xAPI but it is most common to see the specification used in connecting systems together, performing data analysis, and facilitating performance support.

Connecting Systems

For many organizations, getting to a single source of record to account for the learning activity that has occurred across all systems is a dream.  It’s not uncommon to see multiple Learning Management Systems or a host of 3rd party tools be in use across a company. As soon as you have more than one LMS or more than one place in which learners learn, keeping track becomes a headache. And given that the current trend is towards learners learning anywhere and everywhere, that headache is about to turn into a migraine.

Typically, when you have to connect multiple systems together, you end up doing a fair amount of custom work to get them talking to each other. Sales folks for software companies will tell you that their technology will integrate seamlessly because ‘it has an API’. But having just any API is not enough - unless it is a ‘standard’ API, you will still have to write custom integration methods to get systems talking.

There are many relevant standards to be considered. For example, Single Sign-On has a whole raft of ‘standard’ approaches such as SAML or oAuth. But for sharing learning activity data, there is really only one viable API that has been standardized - xAPI. When your systems use xAPI they communicate with each other via the Learning Record Store. All Activity Sources go in, all Activity Consumers come out. Not only does this cut down on the number of custom integrations you need to make, but it also gives you a single source of record for all this activity - the Learning Record Store.

In time this will give you more flexibility and choice in the systems you procure. Right now it can be a very demanding task to migrate between Learning Management Systems. Much of this complexity is caused by learning records being kept inside the LMS, in some proprietary format. By adopting a Learning Record Store you take control of your learning data and keep it separate from any system you might want to replace in the future.

And because Learning Record Stores are interoperable with each other, you can switch out your LRS provider with a minimum of fuss. This is what makes an LRS so different from an LMS - there is a very specific set of rules that account for what an LRS must do. As long as your LRS adheres to these rules, you are safe.

Analysis and Benchmarking

Presuming you’ve got some level of connectivity between systems, the next thing you’ll want to do is analyze how learning takes place across these various systems. Using this data you can draw some conclusions about which forms of training actually impact performance and which learning behaviors are demonstrated by your highest performers.

To do that, you need to track both the training that is viewed and consumed, and the performance data. This is why xAPI should be of interest to people even outside of the L&D department: the L&D team will need access to this performance data if they are ever going to be able to test their hypotheses and to design better learning. Analysis in this style isn’t straightforward and it’s not immediately rewarding. You are stepping towards the world of scientific experiments; attempting to understand the relationship between two variables. And most scientific studies find absolutely no relationship whatsoever!

Science is rigorous. Which also makes it ponderous. If you are wondering why there aren’t hundreds of successful case studies from xAPI that show how training impacts performance, it is not because xAPI doesn’t work. xAPI is simple. It’s just the way we communicate. But actually finding a relationship between two variables in a controlled, repeatable manner is darn near impossible. Just because it’s hard, doesn’t mean you don’t have to try. Imagine using that as an excuse to the board in this age of analytics. “Yeah, well I was going to prove the impact of my training, but it turns out it’s too hard”. That’s a sure-fire way to get your budget cut…

Finally, it’s worth a note about the value of data today for analysis tomorrow. Even if you do not have the ability to analyze data that you are collecting today, you may develop that capability tomorrow. Having a base in historical data will give your analysis more depth and give the opportunity to benchmark current performance against past performance. That’s why we’d recommend organizations get started with xAPI sooner, rather than later. It will take time to build up a meaningful amount of data for use in later analysis.

Performance Support

Whilst you might not get many ‘Holy Grail’ answers from the analysis you will probably get a whole heap of pointers on what works better, and again, xAPI can help here. If you’ve connected your systems together and done the analysis to understand that when a person does X, they may need Y, you can trigger things to happen automatically.

Imagine sending a customized email with revision advice to a person failing an assessment 3 times in a row. Or sending the latest product video out to a salesperson the day before they have a meeting on the new product. Or texting new installation advice out to your on-site engineer right as they turn up on-site. Commercial sites like IFTTT and Zapier have become huge working on just these principles. Now you can do it within the confines of your own organization.

The idea of performing ‘just in time’ interventions whereby the system identifies that you need help right when you need it has been prevalent in the learning industry for a number of years. As we increasingly hear about the role of AI and Machine Learning in technology, you should know that it will be through the standardized use of data that these techniques are actually brought to life. It’s always dull behind the sexiness :)

As a most basic example of Machine Learning, consider a performance support system that not only sends out automated interventions but learns when and how those interventions make a difference to outcomes. If the system sends out help emails and the learner goes on to pass the assessment, it reinforces that the email is a good intervention. If the emails have no effect on the later outcome, then the system learns that this approach has little actual benefit and can try another approach. Right now it might seem like Machine Learning is way away. But xAPI is shaping our ability to exploit these techniques in the near future.


Would you like to learn more about the xAPI the ways it used? Download the eBook The Learning Technology Manager's Guide Τo xAPI and begin your amazing journey.