If you're reading this, you're probably tired of your current LMS. Maybe the admin work takes twice as long as it should. Maybe your participants stopped logging in months ago. Maybe support tickets disappear into a queue and come back a week later with a link to a help article you'd already read.
You're not alone. In 2020, Brandon Hall Group found that 42% of companies were actively looking to replace their LMS.
Most organisations outgrow, or simply grow tired of, their learning platform at some point. And yet many stay far longer than they should, because switching LMS feels like open-heart surgery: years of completion data, compliance certificates, course content, and integrations all live in that system. What if something gets lost on the way over?
That fear is legitimate. But it's also manageable. An LMS migration done well is not a leap of faith. It's a project with known risks, known safeguards, and a clear sequence. This guide walks you through all of it: how to know it's really time, what data to protect, whether to run both systems in parallel, what your new vendor should do for you, and what to do if something goes wrong halfway through.
One thing before we start: you don't have to choose between suffering through another year with the old system and a six-month overhaul. That's a false choice, usually offered by vendors whose platforms genuinely do take six months to set up. Scope the migration honestly and get real help from the new vendor, and it can be done in weeks.
What are the signs it's time to switch?
Five signals are worth taking seriously, and none of them is about one bad week:
- Your admins work around the system, not in it. If your team exports to spreadsheets to answer basic questions, or keeps a parallel folder structure because finding content in the LMS is hopeless, the platform has already lost.
- Participants avoid it. Low login rates, courses started but never finished, managers chasing completions by email. Before you blame the platform, though, it's worth an honest look at whether the problem is the system or how it's being used. We've written about that exact question here: why your LMS isn't being used. If you've fixed the content, the communication, and the manager involvement and people still bounce off the interface, the platform is the problem.
- Support has become a black hole. You submit tickets, wait, escalate, wait again. When you can't get a human on the line for something business-critical, you're carrying risk you didn't sign up for.
- The pricing no longer matches your reality. You're paying for seats nobody uses, or hitting surprise fees for features that should be standard.
- The roadmap has gone quiet. No meaningful updates in a year or two usually means your vendor's attention is elsewhere.
None of this is new, by the way. As far back as 2016, a Brandon Hall Group survey found 44% of companies with learning technology looking to replace their solution, and the top reason, cited by 87%, was the need for a better user experience.
If three or more of these sound familiar, stop asking "should we switch?" and start asking "how do we switch safely?" And when you get to selection, build in the protection you didn't have last time. A structured requirements process, including exit clauses that guarantee you can get your data out, saves you from ending up in this exact spot again in three years. We've covered how to run that process here: LMS RFP template and requirements checklist.
Your LMS migration checklist: know your risks before you move
Before you look at a single new platform demo, map what you're actually carrying. Most migration horror stories come from a dependency nobody discovered until after the switch. Work through this checklist first; everything else in this article builds on it.
The core questions the checklist covers:
- Data: What must move, what can be archived, what can be left behind? User accounts, enrolment history, completion records, certificates, assessment results.
- Content: Which courses are actively used? In what formats do they exist, and can you export them in a standard format like SCORM, xAPI, or cmi5?
- Integrations: What talks to your LMS today? HR systems, single sign-on, calendar tools, reporting pipelines. Each one is a migration task of its own.
- Compliance: Which records are you legally required to retain, and for how long? This is where certificates and completion history become high-risk data (more on that below).
- People: Who needs to be informed, trained, and won over? Admins, managers, participants, IT.
- Timing: Are there training deadlines, audits, or onboarding waves you must not disrupt?
Answer these honestly and you've already done more preparation than most organisations that end up writing angry posts about their migration.
Parallel running or big bang: which migration approach fits you?
You have two basic strategies, and the choice between them shapes your entire risk profile: run the old and new systems side by side for a transition period, or pick a date and switch everything at once.
Neither is universally right.
| Parallel running | Big bang | |
|---|---|---|
| Risk profile | Spread out over the transition | Concentrated on one day |
| Cost during the transition | Double licences during the overlap | One licence, no overlap |
| Admin workload | Two systems to maintain | One system, one push |
| For participants | Can be unclear which system to use | One clear switch |
| If something breaks | The old system is still your fallback | You fix forward, under pressure |
| Fits you when | Long courses that cannot be interrupted, strict compliance deadlines, a large participant base | A compact catalogue, low ongoing activity, a migration you validated first |
Parallel running means participants finish ongoing courses in the old system while new activity starts in the new one. Your safety net is built in: if something breaks, the old system is still live. The costs are real, though. You pay for two licences during the overlap period, your admins maintain two systems, and participants can get confused about where to go. Parallel running fits you if you have long-running courses that can't be interrupted, strict compliance deadlines, or a large participant base where a single bad launch day would be expensive.
Big bang means you migrate, validate, and cut over on a set date. It's cleaner, cheaper, and forces decisions instead of letting them drag. The risk is concentrated: if something is wrong, everyone feels it at once. Big bang fits you if your course catalogue is compact, your ongoing activity is low (summer and year-end are popular for a reason), and you've validated your data migration thoroughly before the switch.
A practical middle path many organisations land on: big bang for the platform, parallel access to the old system in read-only mode for a few weeks. You get one clear "we've moved" moment, but historical data stays reachable while you verify everything landed correctly. Ask your outgoing vendor what a read-only wind-down period costs before you assume it's possible; it's a negotiation point, and one more reason exit terms belong in your contract from day one.
What actually moves with you: data, certificates, content, integrations
Break your inventory into four categories, because each moves differently. "We'll migrate everything" is what makes the project grow beyond its estimate.
User data and history. Accounts, roles, group structures, enrolment records, completion data. The structure rarely maps one-to-one between platforms, so expect a mapping exercise: what your old system calls a "learning path" might be a "course program" in the new one. This mapping is where errors hide, and it's work your new vendor should do with you, not delegate to a template.
Certificates and completion history: your high-risk data. If your training carries compliance weight (safety certifications, regulatory training, licences to operate), the records proving completion are the most valuable thing in your old LMS. Losing them can mean employees having to redo mandatory training, or your organisation failing an audit it would otherwise have passed. Treat this data with its own migration track: export it early, in a format an auditor would accept (including dates, versions, and who completed what), and verify it independently of the main migration. Even if the new platform can't import every historical certificate natively, a complete, well-organised archive protects you.
Course content. Content built in standard formats travels well; ask both vendors explicitly about SCORM, xAPI, and cmi5 support for export and import, and test with a real course before committing. Content built in your old vendor's proprietary authoring tool often doesn't travel at all, and this is where migrations outgrow their estimate. A migration is also a rare chance to leave dead content behind, so use it. If a course hasn't been touched in two years, ask whether it deserves a second life or a dignified archive.
Integrations. Every system connected to your LMS needs to be reconnected, reconfigured, or retired. List them, assign an owner to each, and test them in the new environment before launch.
While you're mapping data, ask the storage question too: where will your data physically live after the move? For European organisations, data residency is increasingly a hard requirement. It's one of the reasons we at Learnifier keep data on EU servers, and let customers choose Swedish hosting if their data should stay in Sweden. Whichever vendor you choose, get the answer in writing.
How do you avoid data loss and downtime during an LMS migration?
Five habits, all about copies and verification, cover most of the data-loss risk:
- Take a full export before anything else. Before any migration work starts, pull a complete export of users, completions, certificates, and content from the old system and store it somewhere you control. This snapshot is your insurance policy and the foundation of any rollback (more on that later). Do it while your old contract still guarantees you access.
- Migrate in stages, verify each stage. Users first, then history, then content, then integrations. After each stage, check counts and spot-check records: does the number of completions in the new system match the export? Do ten randomly chosen participants have the right history? Small verification loops catch errors while they're cheap to fix.
- Freeze changes during the final migration window. Pick a short window, communicate it clearly, and pause admin changes in the old system so you're not migrating a moving target. Note that this doesn't have to mean downtime for participants; with a staged approach, the actual unavailable window is often a matter of hours.
- Protect ongoing courses. Nobody should lose progress halfway through a certification because of your timeline. Either let ongoing cohorts finish in the old system (the parallel-running argument) or schedule the cutover for a natural low point in activity.
- Keep the old system in read-only mode as long as your budget allows. It's the cheapest verification tool you'll ever have: any dispute about whether a record migrated correctly is settled by looking at the source.
None of this is technically hard. What it takes is sequencing and a new vendor who has done it before. The harder part is people.
Change management: how do you bring your people along?
A technically flawless migration can still fail on launch day if nobody wants to log in. The system switched; the people didn't.
Switching LMS is an easier change to sell than most IT projects, because you're usually replacing something people already dislike. Use that. Don't frame the migration as "we're changing systems" (nobody cares). Tell people what stops being annoying: fewer clicks to find your course, certificates you can actually download, a platform that works on your phone.
A few things that consistently make the difference:
- Tell admins first and involve them early. They know where the bodies are buried in the old system, and they'll be answering everyone else's questions. A migration designed without your power users will miss things only they know about.
- Communicate dates people can plan around. "Finish ongoing courses by the 15th. New platform opens the 20th. Old system readable until month-end." Three lines are enough.
- Give managers something to say. Participants listen to their own manager more than to a project mailbox. A short message managers can forward, plus one talking point for their team meeting, travels further than any launch campaign.
- Make the first login rewarding. The first thing people see in the new platform should be relevant to them: their course, their history, their next step. An empty dashboard on day one confirms every sceptic's suspicion.
And keep expectations honest. Some things will be different without being better in week one. Say so up front. It costs nothing, and people are more patient with what they were warned about.
What support should you demand from your new vendor?
Too few buyers ask the vendor directly: what exactly will you do for us during the migration, and what does it cost?
The answers vary wildly across the market, and they tell you a lot about who you're about to marry. Migration support falls into three categories.
- What should be free: a named contact who is responsible for your onboarding (not a ticket queue), help mapping your data structure to the new platform, guidance on export formats from your old system, admin training, and honest answers about what won't transfer cleanly. If a vendor charges extra for basic onboarding help, that's a preview of the relationship.
- What's reasonable to pay for: hands-on bulk migration of large content libraries, custom integration development, rebuilding courses trapped in a proprietary format, historical data transformation beyond standard imports. This is genuine project work; paying for it is fair, but the scope and price belong in your contract before you sign, not as a surprise in month two.
- What's a red flag: "our documentation covers all of that," no named person, migration services only through a third-party partner you've never met, or an onboarding process that assumes you have a project manager to spare.
The vendor's behaviour during your migration is the most honest preview you'll get of their support afterward. A vendor who is responsive, human, and generous while winning you over will usually stay that way; one who is slow before you've even signed will not speed up later. This is why we at Learnifier put people, with names and phone numbers, at the centre of onboarding: the weeks when you're moving your data and rebuilding your setup are exactly when you need someone who picks up the phone.
So ask the uncomfortable questions in the sales process. Who, by name, will help us migrate? What did your last migration of our size actually take, in weeks? What's included, what's billed? Write the answers into the agreement.
What are the most common LMS migration mistakes?
Most failed migrations fail in the same few ways. In rough order of frequency:
- Migrating everything. Moving ten years of obsolete courses and inactive accounts multiplies the work and imports the clutter that made the old system unbearable. Migrate what's alive; archive the rest.
- Skipping the test migration. Running the real migration as the first migration means discovering mapping errors in production. Always do a trial run with a subset of real data and verify it before the full move.
- Treating certificates as an afterthought. Compliance history gets bundled into "data" and nobody checks it until an audit does. Give it its own track, its own verification, its own archive.
- Underestimating integrations. The LMS migrates fine; the SSO doesn't, and nobody can log in on day one. Test every connection in the new environment before cutover.
- Announcing before validating. Telling the whole organisation "we're live!" and then finding broken data is how you burn trust you'll need for months.
- Having no rollback plan. This is the mistake almost no migration guide talks about, so let's fix that now.
Your rollback plan: what if it goes wrong halfway?
A rollback plan is what lets you proceed with confidence. It has four parts:
- A restore point. The full export you took before starting (see above) plus the old system kept live or read-only through the transition. As long as those exist, no mistake is fatal.
- Defined triggers. Decide in advance what would make you pause or reverse: completion data that doesn't reconcile, login failures above a threshold, a compliance record you can't verify. It's a much easier call to make in advance.
- A decision owner. One named person with the authority to say "we stop, we fix, we retry next week." Rollbacks fail when that call has to go through a committee.
- A communication template. Two pre-written messages: "we're delaying the switch, here's why and when" and "we're back on the old system temporarily." With those two written, a delay is just an email.
In practice, a rollback rarely means abandoning the migration. It usually means reverting to the old system for days while you fix a mapping error, then cutting over again. Organisations with a rollback plan treat that as a hiccup. Organisations without one treat it as a catastrophe, because by then they've often let their old contract lapse. Time your old system's termination for after successful validation.
After the migration: validation and the first weeks
Two things close the migration: the numbers reconcile, and people are actually working in the new platform.
Validate before you celebrate. Reconcile the numbers: user counts, completion totals, certificate counts against your pre-migration export. Spot-check real records, and let a handful of admins and managers verify their own teams' data; they'll catch things no script will. Test the critical journeys end to end: log in through SSO, enrol, complete a course, download a certificate, run the report your leadership asks for every quarter.
Then watch the human signals. Login rates in week one and week three. Where support questions cluster (they'll show you exactly what needs a better guide or a five-minute walkthrough). Whether managers can find their teams' progress without asking. Fix the small frictions fast; early annoyances harden into "the new system is also bad" faster than you'd think.
And keep one thing in view: the migration got your data safely into a new home, but it hasn't yet built new habits, launched new programs or delivered the outcomes you switched for. After the migration, a new implementation begins, and that's a project with its own playbook: LMS implementation checklist.
The fear that keeps organisations locked into an LMS they've outgrown is almost always bigger than the actual risk, once the risk is named, listed, and planned for. A good migration is a named list of risks, worked through in order, with real help from the vendor you're moving to. If you want to zoom out first and revisit what a learning platform should be doing for you in the first place, do that before you start. Then take the checklist above, walk through it with your team, and find out how much smaller this project is than it looks from the outside.
LMS migration FAQ
How do I know it's time to switch LMS?
Look for structural signals, not bad days: admins working around the system in spreadsheets, participants avoiding the platform, support tickets going unanswered, surprise fees, and a roadmap that's gone quiet. If several of these persist after you've fixed content and communication on your side, the platform is the problem and it's time to plan a switch.
Do you lose data when switching LMS?
No, not if you plan for it. Take a full export of users, completions, certificates, and content before migration starts, migrate in stages, and verify counts after each stage. Keep the old system in read-only mode during the transition as a reference. Data loss in LMS migrations comes from skipped verification, not from the switch itself.
How long does an LMS migration take?
It depends on your data volume, content formats, and integrations, but it does not have to take six months. With a clean inventory, standard content formats, and hands-on vendor support, many organisations complete the move in a few weeks. The biggest time drivers are content trapped in proprietary formats and integrations that need rebuilding, so scope those first.
Can you run the old and new LMS in parallel during the transition?
Yes, and it's a common risk-reduction strategy. Participants finish ongoing courses in the old system while new activity starts in the new one. The trade-off is paying for two licences and extra admin effort during the overlap. A popular middle path: switch fully to the new platform but keep the old system in read-only mode for a few weeks while you verify the migrated data.
What happens to old certificates and completion history?
They should be treated as your highest-risk data. Export them early, in an audit-ready format with dates and completion details, and verify them separately from the rest of the migration. Even if the new platform can't import every historical record natively, a complete external archive protects your compliance position and spares employees from redoing mandatory training.
What support should the new vendor provide during migration?
At minimum, and at no extra cost: a named contact who owns your onboarding, help mapping your data to the new platform, guidance on exports from your old system, and admin training. Hands-on bulk migration of large libraries or custom integration work is reasonable to pay for, but scope and price should be agreed in the contract before you sign.








.webp)

.jpg)





