You Can Get There From Here: The xAPI Learner Personalization Framework

By Myra Travin (Educational Futurist)

“The simplicity that precedes complexity is useless; the simplicity that follows complexity is the pearl of great price.” — Marvin Anderson

One of the greatest obstacles to the implementation of xAPI in organizations is the tension between the need for a turnkey solution and a highly personalized learning experience environment. In my xAPI Camp session held at Amazon in July, I introduced the concept of a personalized learning framework—an application that recognizes individual learner performance data and predictability and places it within a framework for implementation.

If you want to meet the diverse needs of both learners and the organizations that support them, it is critical to understand that the needs of each are situated on diametrically opposed sides of each other on an axis of execution. The needs of the organization are for evaluation and data that shows return on investment—principal goals of the business. The goals for an individual learner stand in stark contrast to the organization’s need for order, economy, and collective application.

When mobile adoption and access surpassed 2.8 billion devices in 2014, there was no doubt in the minds of learning experience (LX) designers that learner choice and access would supplant current methods of learning and development. The question of what would replace eLearning was still to be realized. The groups that adopted distance learning and SCORM were meeting the needs of the business for value and assessment, but the anarchism of mobile personalization made those efforts obsolete. It created opportunities for predictive performance support data, but required something that could accommodate mobile and offline uses to track it. xAPI was born in between a vision of possibilities and the reality of a lag in commercial adoption for both tools and resources. This, to a certain extent, is where things remain today: a great idea in search of a plan of enactment.

In this space, the ideas which served us in the past—straightforward tools fashioned to assist learning and development professionals to create content without the need for coding—do not meet the requirements of learner-controlled environments, data management, and business analytics. New tools have been and are being created to meet the shift to mobile access, but often they are still a simplified view of the same notion: that a single collective solution will be sufficient.

I suggest that it is not another tool we need, but a strategy that will intersect both learner and business goals into a way forward that is not simplistic, but achievable. In that way, each side of the equation impacts the other and the goals of both are attained.

My design for the Learning Personalization Framework (LPF) is a roadmap that takes a step-by-step approach to incorporate the shift to learner control, mobile and cloud learning, and integrates xAPI for assessment and predictability. The framework has six steps for application:

  1. Learner Evaluation and Needs Assessment
  2. Educator-created LX and Program and Platform Design
  3. Learner Control Mechanisms and Tools
  4. Relationship Strategies
  5. Group and Community Involvement
  6. Data Conjoint Analysis and Iteration

The purpose and focus of each step is outlined below:

  1.      Learner Evaluation and Needs Assessment

In the development framework, the learner assessment stage begins the process of implementation. The needs assessment phase of the framework evaluates learners within the context of skill requirements, and the diagnostic data that is gathered sets the design parameters for progress within the model. The model exists to assist organizations to move from a technology-dependent, instructor-controlled environment to a learner-choice-driven growth engine which supports the belief in self-directed algorithmic constructivism.

  1.      Educator-created LX and Program and Platform Design

At the launch of the model, the foundational aspects of all areas of course design, technology platforms, applications, and resources must be put into place to create the opportunities for access and communication that will be afforded in the later stages. As this model devises learner involvement to begin in the third stage, the resources are thus seen as static at this stage. All aspects of instructional content design will be located at this stage of the model, and the stratagem of flipped classroom access and full content access is critical to the outcomes of this stage. It is to be noted that this model is able to accommodate numerous instructional design approaches, and does not rule out any theoretical approach.

  1.      Learner Control Mechanisms and Tools

With all content, platforms, and communication protocols in place, the model moves to the third phase—introduction to the concept of learner choice. If all curriculum is in place at the previous stage, the emphasis now shifts to point-of-need access by individual learners to resources, course materials, and tools. Learners will have constant access (how and when) to all resources. This access will accelerate competencies and knowledge acquisition and provide a stimulus for dynamic stages of the model to follow.

  1.      Relationship Strategies

The key elements of this stage center on the introduction of partnership and support relationships of experts to learners. Technology that allows for interaction and communication are critical to the success of this stage. The foundations for this stage are in Stage 1 but implemented in Stage 4, as the model moves into this relational performance support aspect. As suggested, even though this is not employed in the earlier stages, it is critical to have both technology and strategy placeholders in the design.

  1.      Group and Community Involvement

As the last of the framework stages before analysis, the greatest requirement of choice, learner access, interactivity, communication, and performance support is evidenced. This stage is the fully realized learner ecology and has the most sophisticated requirements. It is clear that not every environment will be fully realized (based on the outcomes of the diagnostic), and it may take several iterations for the design to include the more complex requirements. As in other growth cycles in nature, the system itself will grow in complexity until it reaches a threshold of a desired state.

  1.      Data Conjoint Analysis and Iteration

The outcomes of the summative final data analysis will guide the return back to Stage 1, and indicate areas of advancement and expansion of the original design hypothesis. Once this data has been incorporated into the content, usage, and design, the ecology is in ever-increasing cycles of health and growth.

Recommendations for Final Iteration

The purpose of this stage is to determine if all factors—design, technology, data gathering, and implementation—have led to an increasingly better environment for the learner. This final stage is the summative evaluation of all of the pieces of the model, and it will suggest changes in the approach that will better support the viability of the design.

Throughout the framework, each step builds upon the next and provides a blueprint for the creation of a xAPI personalized learning environment. This process is built upon the Learner Personalization Diagnostic (LPD) tool that uses learner pre-evaluation data to suggest the design requirements needed for implementation. The tools are not static, nor recommended and limiting, and one can begin at their own company learning strategy and configuration. This means that you do not have to scrap your current process and tools, but update and integrate them into a new approach to meet the needs of new learners and goals for more data-driven and dynamic learning environments.


In the diagram, you can see a snapshot of visualized data which diagnoses the state of the current learning environment for a global implementation of a Learning and Development enterprise. The straightforward visualization shows that this learner group is highly relational and the tools that would support communication and performance are subscribed to propel the learners into a more active learning ecology environment.  

Experience API statements will be defined, based upon this diagnostic, and will suggest next steps within the framework and provide guidance into transitional action verbs and their relationship to objects within context properties. This lays the groundwork for all activities in the tincan.xml file for LMS reporting. This design from data to data approach establishes personalization definitions from the start of the design and delineates the points in the learning process where iteration then redefines the environment, based on learner performance data and outcomes.

Once you have established the underpinning for digital education platforms and tools for your learner group, gleaned from them the types of curated resources they find valuable and their environmental, communication, relationship, and community preferences, and established and integrated the attendant xAPI statements into your design, you will have created a roadmap you can follow now and in future implementations. This is the turnkey simplicity that arises from the original complexity of personalized design considerations which are diagnostic and based upon concrete learner evaluation data. You will then know what to do and how to do it so that it is actionable, repeatable, and data relevant. Finally.

To learn more, please attend the xAPI Camp – Toronto on November 17th. I will be presenting the model at this event.

MYRA TRAVIN (Educational Futurist)

16ef661e45a0-Myra_009In the equation of technology and learner, Myra Travin believes the key element is how humans develop a relationship with the tool they are using and how together they co-create a future state of resourcefulness. She believes that the music is not in the piano, but in the interaction between user and technology. In other words: Learners First. People are the key to innovation, she believes, not simply the systems they use—but systems can reveal the true nature of people through algorithmic predictive data strategies. In her new Learner Personalization Driven Ecology design, she brings all elements of personalization, choice, and access into self-supporting environments. She is an Innovative LX Designer and Educational Futurist with significant experience in user research, use cases, user journeys, surveys, academic papers, data collection, analysis, and user testing within project implementations of Information Security, Organizational Development, Mentorship, UX/UI Design, Predictive Analytics Design, mLearning Design and Delivery, Change Management, Stakeholder Communications Strategies, and CRM projects. Myra has worked in disciplines ranging from higher education with universities to such Fortune 500 companies as Microsoft, Hewlett-Packard, BP, Walgreens, PwC, and SPSS Inc., and public sector agencies such as Los Alamos National Labs, The Bill & Melinda Gates Foundation, and The Ministry of Forests in Canada. She has a deep interest in and a commitment to relationship and interaction within learning environments and assisting individuals in understanding their nature, behavior, and motivations to significantly increase their performance.

Speak Your Mind

This site uses Akismet to reduce spam. Learn how your comment data is processed.