You may have heard of the benefits of xAPI; here's how to put it into practice.

Since its first production release in 2013, Experience API has been gaining traction, enabling L&D professionals to look deeper into the landscape of learning and performance within their organizations. As the specification and the tools and services to support xAPI gain maturity, L&D functions will be looking to begin their own implementation.

One of three distinct approaches for entry provide options for L&D functions seeking to implement xAPI. Case use examples for each of the approaches will help you visualize the possibilities.

xAPI defined

xAPI is a specification for learning and performance data. Whereas SCORM tracks limited data about e-learning completions within a learning management system, xAPI supports in-depth tracking for a variety of learning experiences, even those that take place outside an LMS (see Figure 1). L&D professionals can use xAPI to track interactions with a learning object, activity across an entire learning experience, and on-the-job performance—and then share that data across platforms.

Many L&D functions find that xAPI is a compelling alternative to SCORM, offering both greater depth of analysis and the ability to track data from a wider variety of sources all in one place. xAPI overcomes many of SCORM's limitations—for example, a sole focus on e-learning, lack of support for mobile content, and limited and shallow data—while offering some of SCORM's benefits, namely, a shared industry specification for data interoperability. That interoperability promise means organizations can confidently select tools and platforms that will work together with relative ease, and it opens up the market for software and service providers.

Components of the interoperable xAPI learning ecosystem

The robust data is made possible via an interoperable learning ecosystem that is increasingly easier to implement as software providers expand their xAPI support. Given its roots in SCORM, xAPI is most widely adopted in the e-learning space, with implementation spreading to other elements of the learning data ecosystem.

The ecosystem comprises such components as data providers, learning delivery layers, data stores, and data consumers. Figure 2 outlines a few examples of the components and some of the state-of-the-market limitations and opportunities of using them within one's own ecosystem.

Not surprisingly, the tools and platforms that are closer to an organization's core L&D function are more likely to support xAPI. What's more, L&D has greater influence and leverage with those vendors to push for interoperability. The further out of L&D and into the business a system is, the less likely it will support xAPI. Data in this sphere may need to be adapted from xAPI to other formats to be integrated here.

Three approaches

Companies are taking several different courses as they chart their way in this new territory. The approach they choose depends on the L&D team's knowledge, capabilities, and resources as much as on the organization's commitment to data-driven innovation. When the L&D function partners with the IT and functional departments, it's able to get started with bigger projects than teams with limited budgets and clout in the company.

We're seeing three approaches to getting started: unreleased experiments, small projects, and large projects.

Give it a try: Experimental projects

These unreleased experimental projects are spawned from the imagination and ingenuity of instructional designers and learning technologists and are usually done with the L&D function's support and the expectation to report back with results. These professionals learn about xAPI while implementing early projects, usually with free or inexpensive tools or sometimes writing their own code to get the results they want.

They use the side projects developed in this way as early prototypes or proofs of concept that they can then employ as part of a business case for future projects. If the L&D team releases the projects, a small population of learners uses them on a non­critical business function. Many of these projects never leave pilot mode.

Project goals: Whet the organization's appetite for more and learn something along the way.

Approach: It's OK to break things and fail as you learn because you're not close to real data or mission-critical processes.

How to do it: Plenty of tutorials are available. Use existing e-learning authoring tools and a free trial version of a learning record store (LRS) to get started. More formal learning exists via online learning platforms, vendor hands-on workshops, and the xAPI Learning Cohort.

Examples: xAPI Learning Cohort teams have taken on experimental projects, working in an environment with free tools and teams of volunteers to test various xAPI use cases.

For instance, Team Unity 3D-sandbox tested game-based learning to consider how a Unity 3D scene could work alongside a Storyline e-learning module to create a more interactive learning experience. The team measured time in xAPI-tracked activities within the scene to evaluate and enhance the user experience within the game player context.

Team Alexa used the Alexa Skills within Voiceflow to design an interactive customer support scenario where Alexa acted as the customer and the learner served as a customer support representative tasked with troubleshooting an issue. The team captured xAPI statements directly from Voiceflow and evaluated the learner's interaction with the role-play customer.

Team Adaptive Learning used a WordPress website plug-in to organize and display content to learners based on their responses, with the intended goal of sequencing content using machine learning. To reach the goal, each content piece within the learning experience will need to be tagged with a learning objective and difficulty level and then algorithms applied for sequencing.

Go small: Self-contained projects

Small projects are self-contained learning experiences that take advantage of xAPI without taking on the risk associated with expensive tools or bringing mission-critical functions to a grinding halt if things don't work as planned. Organizations that start with small projects usually have bought into the promise of xAPI but still take a measured implementation approach. Such projects often require the L&D team to work with IT to, at least, vet and integrate new technology purchases (such as an LRS), although the integrations with other systems are generally kept to a minimum.

Project goals: Do it and learn from it. Build the infrastructure and organizational knowledge for a broader implementation.

Approach: Don't break things that people care too much about. Don't spend so much money that you draw too much concern or attention from leadership. These are still small trials.

How to do it: Get a willing project team, a willing project sponsor, and a supportive IT business partner. Choose a project for which SCORM doesn't provide sufficient data. Plan to do a lot of xAPI education with your team, L&D peers, and sponsor. Work with an LRS provider that is willing to offer support as you learn, and plan to pay for it (free trial tiers typically aren't robust enough, and you should not expect extensive vendor support when using them).

Examples: These projects all represent first or early xAPI implementations for the respective organizations, establishing a foundation for future work.

The customer learning team at supply chain modeling software company LLamasoft used xAPI to bypass an LMS for its customer-facing learning videos. This proof of concept enabled the team to test product tutorials hosted within its help system and to understand via interaction data how customers were using them. The process gave great insight into production value and users' content needs.

The Ann Arbor Hands-On Museum set up iPads with interactive content at 20 exhibits throughout one of its facilities. Students visiting the museum wear badges with low-energy Bluetooth beacons to identify them with a particular class, enabling the iPads to serve the appropriate grade-level-specific content. The museum then uses the xAPI data to track exhibit interactions and show quantitative data related to the learning experience and the state science standards to which schools and teachers are held.

A global sales team implemented a multifaceted learning program with five key deliverables that used extensive xAPI data tracking to gain insight into learner activity and performance. The team tested learner performance via question activities, engagement with interactions, audio replays, the use of hint buttons, and text responses (evaluated for effectiveness). Although the insights gave rich data into the learner experience, the team still relied on SCORM in its LMS to track completion data for the critical formal e-learning and testing components.

Go big: High-visibility projects

When the stakes are higher, the organizational support for big projects is often higher too. These larger projects usually involve a partnership and tight coordination among a strong L&D team, IT, the business function, and possibly one or more vendor partners deeply engaged in the effort.

Project goals: Leapfrog an outdated infrastructure and do something the company must do and that absolutely cannot be done with existing data schemes.

Approach: Take on a big mission-critical project that will get a lot of attention and the resources and support that go along with it. Such projects are too big to fail.

How do to it: Similar to a small project, you'll need to assemble a willing project team, a willing project sponsor, and a supportive IT business partner. Generally for projects of this scale, closer collaboration with your learning technology vendor partners and custom software development will be required. Here again, plan for extensive education about xAPI among your team and sponsor; vendors can help. For larger projects such as these, most teams will develop and release in iterations, starting with a small pilot group, gathering data, and refining the experience before a broader implementation.

Examples: These projects are much larger in scale relative to their organization's learning infrastructure and, in some cases, represent the entirety of the learning a company offers to its employees.

An asset management and financial services firm developed a retirement academy that included 18 learning modules across three levels to support customers and staff with retirement planning. The training is asynchronous, and the team wanted to use data to drive curiosity and promote engagement with the content. Thus, it used data to build a bespoke email application that takes a feed from the learning analytics platform and sends learners an email once a month. The email provides learners with a brief overview of how they are progressing in the academy.

Because a large machine manufacturing corporation found that standardizing the customer experience across its dealer network would increase global revenue by $20 billion, the company sought to create a simple, standard approach that all dealers could follow, regardless of their size and business environments. The manufacturer developed a multimodal training program and an interoperable learning ecosystem that received xAPI data from Inkling, OpenSesame, and Kaltura and converted CSV data from QuestionMark and LogicBay to xAPI Data within Watershed. That enabled the company to analyze the impact of learning on the business. Since implementation, the manufacturer has reduced distribution costs to serve customers.

Opportunities ahead

As L&D functions make room for greater xAPI adoption, the approaches above provide three key entry points for your organization. Evaluate what approach, if any, works best for your team and determine what new opportunities xAPI can offer.