A training team that wants to add conversational practice to its catalogue almost always asks IT the same question: “can we package it as SCORM?” The instinct makes sense. SCORM is the format everyone knows, the one most modules arrive in, the one the LMS swallows without argument. It is rarely the right answer here — and not because SCORM is outdated.
SCORM, LTI and xAPI are not three successive versions of the same idea. They are three standards solving three different problems: distributing content, connecting two platforms, describing in detail what a learner actually did. They can be combined. They cannot be swapped for one another without consequence.
This article compares them on one specific case: a conversational role-play that runs live and gets replayed. We will look at what each one genuinely does, why LTI 1.3 is the right vehicle for this kind of content, and why xAPI keeps its relevance where you might not expect it. The full set-up is covered in our guide to integrating Face Up with an LMS.
SCORM is a packaging format. You build a module in an authoring tool, export it as an archive, and import it into the LMS. The content then lives inside the platform, served by it, and talks back to say where the learner stands: started, completed, passed, with what score.
It is a distribution model, and it has served the industry well. It makes content portable between LMSs, it works without a connection to the public internet, and it requires no negotiation between two systems. For an awareness module, a procedure or a quiz, it remains the default choice and there is no reason to move away from it.
LTI carries no content. It establishes a trust relationship between the LMS and an external tool: the LMS knows who it is talking to, the tool knows who is arriving, and the learner moves from one to the other without a second login. The content itself stays with the tool.
The version that matters today is LTI 1.3, also known as LTI Advantage. It brings three services that change everything in our case: Deep Linking to choose what gets launched, AGS to bring results back, NRPS to keep enrolments in sync. We covered them in our explanation of LTI 1.3.
xAPI, sometimes called Tin Can, is a vocabulary. It lets you write structured statements — an actor, a verb, an object — into a learning record store, or LRS. “This learner replayed this scenario”, “this learner paraphrased at this point”, “this session lasted seven minutes”.
xAPI handles neither display, nor authentication, nor launch. It is not a rival to the other two: it is a traceability layer that can sit alongside either of them.
SCORM was built for content you move through. A sequence of screens, a video, an assessment at the end. The path is known in advance, the content is stable, and there is little to report back: where the learner is, whether they finished, what score they got.
On that ground SCORM is excellent and will remain so. If your catalogue is made of modules produced in-house, the question does not even arise. The distinction between these classes of tooling is developed in authoring tool, LMS and LCMS.
This is the main limitation, and it is structural. A SCORM package is a snapshot. The day the scenario changes — a line to adjust, a character's behaviour to revise, a new variation of the situation — you have to re-export the package and re-import it into every course that uses it.
On a stable module, that happens once or twice a year. On a conversational role-play, adjustments are frequent by nature: you refine the scenario precisely by watching how learners play it. Re-importing then becomes a permanent overhead, multiplied by the number of courses involved, with the familiar risk of versions drifting apart from one course to the next.
A conversation does not break down into screens. It produces an exchange, moments where things tip one way or another, a debrief, and often several successive attempts. SCORM can say “completed” and “score”. It cannot cleanly structure several dimensions of assessment, nor handle gracefully the fact that a learner replays the same situation five times in order to improve.
That is not a flaw in SCORM. It is content that does not match its model. Forcing the format produces a poor result on both sides: a fragile integration and reporting nobody can use.
With LTI there is no package. The scenario is hosted and maintained on the Face Up side, and the LMS sends the learner to it. When the scenario changes, the change is immediately live for every course pointing at it — no re-import, no version to resynchronise, no work for the LMS team.
This is the decisive point for this kind of content. A role-play improves continuously: you adjust a reaction, add a tenser variation, fix a phrasing that misleads. That short cycle is only sustainable if updating costs the organisation nothing.
Deep Linking lets the trainer open the scenario catalogue from inside their own course and pick the one they want. No file to request, no link to copy, no ticket to raise. The mechanism is described in choosing your scenario from the LMS.
That autonomy has a concrete teaching effect: the trainer can vary situations from one group to another, or switch scenario mid-programme because they have seen where the difficulty lies. With an imported package, every change goes back through administration.
AGS returns a score out of 100 and completion to the gradebook, and it allows several grade lines for a single activity. One role-play can therefore feed several columns — listening, clarity, holding the line — instead of a single figure that flattens the differences. We develop this in AGS and grade passback for soft skills.
NRPS completes the picture by syncing enrolments: the learners on the course are known to the tool without importing a list. And the learner never leaves their environment — the role-play opens in an iframe inside the LMS, with no second login.
This is xAPI's strength, and nothing else matches it. Where AGS says “68 out of 100, completed”, xAPI can describe the session: how many attempts, how long, which stages were reached, what changed between the first and the last try. For a team that wants to analyse progress in detail, this is the right layer.
That granularity serves two purposes. First, improving the scenarios: seeing where learners drop off tells you far more than an average. Second, looking for effect, when you want to connect practice to a business indicator, as we explain in evaluating and measuring soft skills.
The confusion is common. The LMS gradebook supports day-to-day teaching decisions: who did what, who is falling behind, who has passed. The LRS supports analysis: aggregating, cross-referencing, comparing over time across whole populations.
Two needs, two audiences, two rhythms. A trainer will not query an LRS between two sessions. A programme owner will not run an impact analysis from a gradebook.
LTI and xAPI are not mutually exclusive. LTI carries the launch, the identity and the grade passback; xAPI carries the fine-grained trace to an LRS when the organisation has one. A team without an LRS has no use for xAPI today, and that is fine: the essentials will already be in the LMS.
So the right question is not “LTI or xAPI”, but “do I have an analysis need that goes beyond what my gradebook can show”. If the answer is no, do not build infrastructure nobody will use.
Three criteria settle the great majority of cases.
A mixed catalogue is the normal situation: SCORM modules for stable content, LTI activities for whatever is live and hosted elsewhere. That is the logic set out in Face Up in the training stack.
Before deciding, put a number on one thing: how many courses would use this content, and how many times a year it will change. Multiply. That product is the real cost of choosing SCORM for living content, and it is almost always underestimated because it is diffuse — a few hours here, a ticket there, a forgotten version somewhere else.
On a large rollout this cost becomes the limiting factor: it determines how often you allow yourself to improve your content. We return to that effect of scale in deploying soft skills training at scale.
Before committing, check three things with your platform team: that your LMS genuinely supports LTI 1.3 and not only an earlier version, that Deep Linking and AGS can be enabled, and who holds the rights to register an external tool at site level.
The two most common routes are documented step by step: the Moodle integration and the Rise Up integration. In both cases the work is short once those three points are settled in advance.
SCORM is not the wrong standard: it is an excellent standard for content you move through. A conversation that runs live and gets replayed simply does not fit that model, and forcing it in costs more than it returns.
LTI 1.3 answers this case precisely: live content, hosted and updated by the supplier, chosen by the trainer, with results coming back to the gradebook and a learner who never leaves their LMS. xAPI sits on top when the fine-grained trace has a real use — and only then.
The decision is therefore not ideological. Look at the nature of the content and how often it changes: the right standard names itself. And if your LMS alone is not moving relational skills forward, the reason lies elsewhere — we explain it in why the LMS alone is not enough.
