An LMS launch demo cannot prove SCORM 2004 or xAPI support. Use representative packages and an evidence table to test sequencing, saved attempts and LRS records.
A training course opens in an LMS demo. The video plays, a quiz appears, and the vendor ticks the compatibility box. Before signing, ask the learner to fail the quiz, close the course halfway through remediation, and return tomorrow. Then ask an administrator to explain the saved attempt. That second demonstration tells you much more about whether the proposed system fits your content.
If you are asking which LMS supports SCORM 2004 and Tin Can API, specify what support must accomplish. A product name alone leaves unanswered questions about editions, players, plugins, learner identity and reporting. The practical buying decision is whether a named configuration can run your packages and preserve the evidence your training process needs.
This article proposes an acceptance framework for that decision. It is not a product ranking or a certification claim. For the broader choice between standards, start with our SCORM, xAPI and LTI overview.
Write down the configuration behind the checkbox
Ask the supplier for the LMS release, package standard and edition, content player, authoring-tool export settings, and any external service involved. Record whether the proposed capability is native, supplied through a plugin, or delivered by a separately licensed player. Those are different operating arrangements even when the learner sees the same launch button.
Consider a hypothetical proposal with an LMS, an external content player and a separate LRS. The learner may see one launch button, but the acceptance plan should name all three components. Ask which component enforces the course sequence, which keeps the attempt record and which produces the administrator report. If those responsibilities are unclear, the support checkbox is not specific enough to approve.
For any proposed integration, get the supplier to name the supported LMS and plugin versions in writing. A documentation reference establishes the architecture to investigate; it does not establish that every plugin works with your deployment. Ask who handles upgrades and which support team takes the first ticket when the player and LMS disagree.
Bring packages that expose different failure modes
Use a small, deliberate acceptance set. Select one ordinary package representative of most of your catalogue, one that exercises branching or prerequisites, and one that learners routinely resume. Include an actual export from your authoring tool rather than relying entirely on a vendor's demonstration content.
For SCORM 2004, name the edition. Rustici Software's SCORM versions explanation describes multiple editions and explains that sequencing and navigation govern progression between shareable content objects. A course with one simple learning object will not meaningfully exercise a multi-object prerequisite rule.
Define the expected sequence in ordinary language before testing: a learner must complete the safety lesson before opening the final assessment; an unsuccessful assessment should permit the intended remediation route. These are proposed acceptance cases, not statements that every SCORM course implements those rules. Test the behaviour your package actually declares.
Keep an unchanged copy of each trial package and its export settings. If the supplier edits the package to make the demonstration pass, document the change and decide whether the same work would be needed across your catalogue.
Use an evidence table instead of a feature score
The table below is our proposed buying framework. It is original analysis, not a vendor comparison or a formal conformance test. Its purpose is to make each important claim observable.
| Support claim | Trial to request | Evidence to retain | Decision if unresolved |
|---|---|---|---|
| SCORM 2004 sequencing works | Attempt the final assessment before prerequisites, then complete the allowed route | Package edition, expected route and recorded navigation | Hold acceptance for courses needing that sequence |
| Learners can resume | Exit midway, end the browser session and relaunch with the same learner | Saved location and before-and-after attempt record | Require a fix or disclose restart-only behaviour |
| Scores are usable | Complete a failing attempt followed by a passing attempt | Both attempts and the agreed displayed result | Resolve aggregation before certificates depend on it |
| xAPI reaches the LRS | Complete one named activity and retrieve its stored statement | Actor, activity, statement identifier and receipt evidence | Treat collection as unproven |
| Reports reflect the event | Find the same learner and attempt in the administrator's report | Raw record alongside report output and refresh timing | Define the missing mapping or delay |
| The integration is maintainable | Reproduce one failed launch with the support owner | Diagnostic record, owner and escalation route | Include remediation work in scope before purchase |
Do not average these rows into a reassuring overall score. If certification depends on the final assessment, a broken sequence is a release blocker even when five other rows pass. Classify requirements as essential, optional or outside scope before the demonstration so that tradeoffs remain explicit.
For xAPI, follow the record past the launch button
Tin Can API is another name for xAPI. Rustici's xAPI overview explains that applications send activity statements to a learning record store, or LRS, which may be independent or part of an LMS. This separates two buying questions: can a learner start the experience, and can the required learning evidence be retrieved?
For the trial, agree on one learner identifier and one activity identifier. Complete a known activity, then have the supplier retrieve its statement from the proposed LRS. Compare the actor with the test account. A familiar display name on a dashboard is insufficient if the stored identity cannot be mapped reliably to your learner directory.
Next, request the operational report that training administrators will actually use. Decide how a statement becomes a displayed completion, if that is required. Agree on the reporting delay and show what happens during that delay. Capturing an event and interpreting it for a business rule are separate acceptance steps in this framework.
If offline learning is part of the scope, add an explicit interrupted-connectivity scenario. Ask the supplier to show where unsent events remain, how reconnection is handled, and how the same learner is recognised. Do not infer those implementation details from an xAPI label.
Make retries and attempt policy visible
Consider an illustrative onboarding course. A learner scores 60 on Monday, reviews the lesson, and scores 85 on Tuesday. Your policy might require the latest result, the highest result, or an administrator-reviewed retake. State the policy and inspect the output, rather than assuming the dashboard's number means what you intended.
Repeat the trial with two different test learners. That helps expose accidental sharing of progress or confused identifiers. Keep production learner information out of the exercise; invented accounts and synthetic activity are enough to test the mapping.
Ask what administrators can correct without deleting the audit trail. A package replacement, mistaken enrolment or support-assisted reset should have a defined procedure. Document who can perform it and what evidence remains afterwards. These procedures belong in acceptance because a live training programme will need explanations for disputed results.
Turn the trial into a purchase boundary
Approve the configuration you tested. Attach the accepted package editions, player arrangement, plugin dependencies, report definitions and unresolved limitations to the implementation scope. If a vendor proposes changing the player after acceptance, rerun the affected cases before approving the substitution.
Also request the recurring charges separately: LMS subscription, external player, LRS, storage, support and upgrade work. The point is to expose dependencies, not to assume an external service is inherently a bad choice. A clearly owned integration may suit your requirements better than an unsupported native workaround.
For an education platform with custom enrolment or reporting rules, review the education technology context alongside the content-player requirements. Where the missing piece is an integration or reporting workflow, scope it through custom software development.
Bring your representative packages, expected learner outcomes and proposed architecture to Emerging Stacks. A useful first deliverable is a short acceptance plan that identifies the exact support claim, the observable result and the person responsible for resolving a failure. That gives your team a concrete basis for approving the LMS configuration.
Ready to work with us?
Get in Touch


