Canvas alternatives for institutions that have outgrown the academic term

Canvas is built around a term with a start date and an end date, a roster that arrives from a student information system, and a course that closes with a grade. Continuing education, corporate partnerships, a commercial training arm and client academies have none of those things: enrolment is rolling, the cohort is a company, the finish line is a certificate, and the people involved are not registered students. Institutions in that position usually do not need a better LMS. They need a second shape.

What the term-and-roster model looks like on a Tuesday

The model is invisible until you break one of its assumptions. There are about five, and each one is reasonable:

• Enrolment arrives in a batch, before the course opens, from a system of record you do not administer.
• Every course has an end date, and the end date does the cleanup — access lapses, the gradebook closes, the section is archived.
• A person is a student of the institution first and a participant in a course second.
• Progress is a grade against a syllabus, and that grade's destination is a transcript.
• Content is copied forward term by term, so "the course" is really a series of near-identical copies with a year in the name.

Now change one input. A facilities manager at a client company signs up on a Tuesday afternoon for a six-hour compliance course. She needs access within the minute, not on the first Monday of the semester. She will do two modules that night and the rest over three weeks. Nobody will grade her; she needs a certificate her employer's auditor can verify. She is not in the registry and never will be.

Running that inside a term-shaped system is possible, and plenty of institutions do it. You create a term that never ends and a course shell that stays permanently open, so the gradebook never closes and the archive job never fires. You enrol people by hand, or with a script on a server that one person maintains. You keep a spreadsheet of who started when, because the reporting is oriented around terms and yours does not end. You duplicate the course per client, because the client's name has to appear somewhere and there is nowhere else to put it.

Each workaround is small and each one works. The real cost is that after two years the workarounds are the operating model, and the person who knows how they fit together has become load-bearing.

The four shapes that do not fit

Continuing education and professional development

Rolling intake, short courses, learners who are not matriculated, employers paying rather than students, and completion expressed as a certificate rather than credit. The reporting question changes shape too: it stops being "what grade did this student earn" and becomes "how many of this employer's staff finished the mandatory module before the audit date".

Corporate partnerships

A university sells a leadership programme to a bank. The bank's HR team wants to see its own people and nobody else's, wants its brand present, wants a monthly report, and wants staff added from its own HR system rather than by your registrar. Those are tenancy questions, and no amount of course configuration answers a tenancy question.

A commercial training arm

Separate P&L, separate pricing, invoicing and VAT, and a need to move faster than the academic change-control calendar. The arm usually wants its own brand rather than the university's, because it is selling to procurement teams comparing it against commercial training providers.

Client academies

The shape that breaks term-based platforms hardest. A corporate academy running training for 15 client companies from one tenant needs each client to have its own branding, its own administrators, its own learner list, and no sight of any other client. That is not fifteen courses. It is fifteen organisations that happen to share an operator.

Theming is not white-labelling

This is where evaluations go wrong, because both words get a yes in a sales call. Theming means colours, a logo, fonts and some CSS. White-labelling means a person can use the platform for a year and never meet a name they did not expect. To tell them apart, stop asking about branding and list the places a brand actually appears:

• The URL in the address bar while a learner is halfway through a lesson.
• The URL on the password-reset link — the one people paste into support tickets.
• The From address and the sending domain on every notification.
• The login page a client's employee lands on at 8am, cold, after a manager sent them a link.
• The certificate, and the page its verification link resolves to when an auditor clicks it.
• The browser tab title and the favicon.
• What a client's administrator sees when they sign in: their people, or a list that includes everyone else's.

Theming changes the sixth item and part of the first. The rest are architecture — domains, TLS certificates, mail sending, identity, and how tenant isolation is enforced. A platform either had those in the plan from the start or it is retrofitting them, and retrofitting shows: the answers come back with conditions attached.

On Canvas, white-labelling is limited to theming: it is not designed to disappear behind your own brand and domain for separate client organisations. That is a scope decision rather than a defect, and it is set out with sources on our Canvas alternatives page.

The honest case for staying on Canvas

Replacing an LMS at an institution is not a software project with a change-management component. It is a change-management project with a software component, and the software is the easy part.

Consider what is actually built around the thing you are proposing to remove:

• The student information system integration, and the nightly job nobody has needed to touch in four years.
• The LTI tools hanging off it — proctoring, plagiarism checking, publisher content, reading lists, lecture capture.
• Faculty who know where the gradebook settings live and have opinions about them. Retraining them costs a semester of goodwill you may need for something else.
• An accessibility review already done, and procurement paperwork already signed.
• A term rollover that runs itself, and institutional policy written with this platform's vocabulary in it.
• Years of course content, some of it authored by people who have left.

The decision test is not "does the new platform do more". It is: which population is growing, and which one pays for the building? If credit-bearing programmes are the bulk of your volume and the non-term work is a promising side line, migrating everything to serve the side line is a bad trade. Run the side line somewhere else.

Two systems is a normal architecture, not an admission of failure. The registrar keeps the system of record for credit; a second platform carries everything that is not credit-bearing. The boundary is unusually clean here because the populations barely overlap — the client's staff in your compliance course were never going to appear in your student registry.

Be straight about what the second system costs: two support queues, two accessibility postures, two data processing agreements, and a real identity question for the few people who exist in both. Budget for that at the start rather than finding it in month four.

Replacing rather than adding makes sense in three cases: the non-term work has become the majority of your activity; a commercial arm is separating into its own legal entity; or you are at a renewal, the switching cost is being paid in some form regardless, and the roadmap for the non-term work is concrete rather than aspirational.

What to test in a trial, whichever direction you go

Demos are built to succeed. These seven checks are quick and they fail loudly when something is not there.

1. Enrol one person on a Tuesday afternoon into a course with no end date. Come back the following Monday and look at their access, their progress and their reporting row.
2. Put the same email address in two client organisations. Find out whether that is one person with two memberships or two accounts heading for a collision.
3. Give a client administrator an account, then try to make them see another client's learner. If a database rule stops you, good. If it takes remembering not to click something, that is not isolation.
4. Send a notification and read the headers. Look at the sending domain, not the display name.
5. Issue a certificate and open its verification link from a phone on mobile data. See whose name is on the page.
6. Ask for a full data export before you sign, not after — content, learner records, results, audit history, and in what format.
7. Ask the standards question plainly. SCORM 1.2, SCORM 2004, xAPI, LTI 1.3: which ship today, at what version, and what is planned? Get it in writing with a date.

Where Lurno fits, and where it does not

Lurno is the platform we build, so read this section as what it is. It was designed for the second shape rather than the first: nested sub-organisations, each with its own branding, administrators and learners, plus an internal branch tree where a role granted at a branch cascades down it. Each organisation can run on its own custom domain with automatic TLS and send email from its own domain, and certificates carry a public verification page. Tenant isolation is enforced in the database rather than only in the application — separation enforced underneath the platform — which is what makes the client-administrator test above pass by design. There is more on white-labelling and on the institutional view at higher education.

What it is not, stated plainly, because a procurement team will check. SCORM 1.2, SCORM 2004, xAPI and LTI 1.3 are modelled in the product but the runtime is still in development — if standards conformance gates your rollout, that is a conversation and not a tick in a box. Payments and checkout are designed, not built. Self-serve signup is in development; accounts are created by invitation today. Single sign-on today means partner sign-on via signed assertion, not SAML or OIDC; SAML/OIDC, MFA and passkeys are on the roadmap and none of them is available now.

The largest limitation is the one this post has argued throughout: Lurno is not a replacement for a registrar's system of record on a campus running degree programmes. If Canvas holds your terms, your sections and your transcripts, it is doing a job worth keeping. The question is what should carry the work that never had a term in the first place. A sourced side-by-side is on the Canvas comparison page; to run the seven checks against a real tenant rather than a slide deck, book a demo.