Top 10 Things That You Are Going to Mess Up in xAPI: Starting with the Technology

This is the first in a series of ten articles that intend to surface common issues, misconceptions, and ill-advised strategies which have the effect of derailing xAPI implementations. We also attempt to provide some useful ways of approaching an implementation and hope this will prove beneficial. Enjoy the series and feel free to reach out with your own examples and experiences of when things went sideways.

First up: Starting with the Technology

It makes sense that when looking for a technology solution, you’d start by looking at technologies. But xAPI is — in many ways — less about technology than it is about process, design, and problem solving.

The core technology necessarily present in any xAPI implementation is the LRS — the Learning Record Store. Whether implemented as a single monolithic LRS or as a network of federated LRSs, the core purpose is the same: to validate the conformance of data to the xAPI specification (now standardized as IEEE 9274.1.1) and to provide for access to and retrieval of that data. There are a variety of bells and whistles that are available in different LRSs on the market — from front-facing things like dashboards to heavy-lifting type things on the backend related to database management and security. But the core purpose of the LRS remains the same — it’s not a data visualization tool, it’s not an identity management tool, it’s a data validation tool.

With this in mind, consider what happens in an organization that decides to initiate an xAPI project by starting with the technology.

Executive: “We need to modernize. Get xAPI in here, pronto!”

Project Manager: “We need xAPI. How do we do that?”

Dev Team: “I guess we need an LRS.”

Project Manager: “Ok, so get us an LRS.”

Six months later…

Executive: “So, have we modernized? Did we get xAPI?”

Project Manager: “Yes, we got an LRS.”

Executive: “That’s great. What does it do?”

Project Manager: “We have no idea.”

An xAPI implementation must begin by identifying the problem that you are trying to solve. “Modernization” and “Innovation” are not problems. So, what is a problem? And what are the kinds of problems that are worthy of considering an xAPI implementation to solve?

Things like:

“We need to correlate the experiences that trainees are having in our XR environment with the outcomes of the assessments they are completing in the coursework on our LMS so that we can show the ROI and value-add of XR-enabled training.”

“We need an auditable record of learning experiences to measure where the disconnect is between our training modules and on-the-job compliance.”

“While completion data gets us part of the way, we need to better understand the learning activity pathways that led to completion or non-completion so that our ISD team can improve our L&D content and make it more accessible and valuable to everyone in the firm.”

Once you’ve identified the problem, here’s a back-of-the-napkin approach to what you’d need to do in order to implement a solution:

• Take an inventory of your current data availability and identify where there are gaps between the data that you have and the data that you’d need in order to solve your problem

• Talk to IT about the challenges of getting access to or integrating with current or future data feeds

• Based on your inventory and conversations, draw up a table of all of the data available to you in the near, mid, and longer term

• Now, consider your problem and identify the ideal types of activities that learners would experience in the environments available to you which might produce data that would help you to solve your problem

• Compare that model of ideal learner activity with your data availability and your team’s bandwidth and capability to either work with what you have or to create new things

• Once you’ve come to consensus on your capabilities and timelines, start to design an idealized profile of what the data output of this activity might look like under ideal conditions

• Compare that idealized data profile to your reality near, mid, and longer term with regard to needs — especially in terms of technical resources, people capabilities, potential outside partners, expected budgets, ethical or cultural concerns, and what measurable results would make a real impact for your business — and then draw up the tradeoffs

• Debate with your team with a focus on nailing down what’s within the realm of the possible for your organization in the here-and-now — again concentrate your effort on maximizing the value to the organization (even if the use case itself seems relatively small from the 30,000 foot view)

• Write up a proof-of-concept strategy to include a) the description of the learning experiences that you are going to track and the personae of the learners that you’ll be tracking, b) an identification of integrations that will be made with existing systems (pay attention to how systems share IDs and how you would manage learner identification downstream), c) identifications of what new systems will either be built or bought in order to produce, validate, or consume the necessary xAPI (or related) data, d) a description of the expected measurable outcome of setting up this new system and a clear statement of the value-add of setting this system up for the organization, the people, the business, the mission

• Author the corresponding xAPI Profile(s), instrument the data sources, implement one or more LRSs, integrate and connect the systems, evaluate the outcomes and iterate as necessary

Note that there are no hard and fast rules regarding what your xAPI data can do. Some organizations primarily use it to support reporting and transcripts. Others store it for compliance audits. Still others use it to provide evidence to competency assertion systems. And some organizations recently have been using it to create a bridge between AI-enabled systems such as intelligent tutors and business systems such as are used for workforce or mission planning.

Also note that every organization is different — and there may be dozens of competing use cases within even one organization. So, you are not going to find a simple off-the-shelf solution to your problem. There is no dashboard that you are going to be able to hook up to your data that will magically tell you everything you need to know. The truth of the matter is that xAPI is a design tool and it will provide you with a way to manage and iterate the design of your data ecosystem now and over time.

So, don’t start out by thinking about “Which LRS is going to solve my problem?” Start out by clearly articulating what your problem is and then identifying those technical, cultural, and procedural challenges that may exist in your organization that will need to be addressed in order to produce the best outcome. Don’t start with technology. Start with problem solving.

Need help? We live and breathe this stuff. And we’d be happy to help you to solve problems. Feel free to reach out anytime.