We build custom eLearning programmes. So it is worth saying plainly, before anything else, that a great many organisations asking this question should go and buy a licence instead.
Every article you will find on this question was written by a company that sells custom eLearning. Search the phrase and count them. They will each walk you through a fair-looking comparison, acknowledge that off-the-shelf content has its place, and then arrive, with an air of reluctance, at the conclusion that your situation is probably the special one.
This is also an article written by a company that sells custom eLearning. So let us start at the other end.
A note on where this sits
Last week we published a piece on how to choose a custom eLearning company. It answers a question that comes second: you have decided to build something, and now you need to work out who should build it, what a fair quote looks like, and which promises are worth anything.
This article is the question that comes first, and it is the one more people should ask before they get as far as talking to anybody.
Should you build at all?
The two pieces are meant to be read in that order, and a good number of readers should stop after this one.
Three cases where you should buy off the shelf, and not call us
Your content is genuinely universal
Fire safety. Manual handling. Basic cyber hygiene. GDPR fundamentals. Anti-bribery. If a competent generic course on the topic would do the job, a generic course is the correct purchase, and paying to have one built from scratch is a straightforward waste of money. The test is not whether your organisation is distinctive. Every organisation feels distinctive. The test is whether the content is, and for a great many compliance topics it simply is not.
Nothing meaningful happens if nobody applies it
This one is uncomfortable, and it is the most useful question in the whole decision. If a learner finishes the course, forgets all of it by Thursday, and no harm follows, then you are not buying behaviour change. You are buying a record that the training was provided. Off-the-shelf content produces that record perfectly well and at a fraction of the price.
Your headcount is small or volatile
Per-seat licensing is a rental model, and rental models are exactly right when you cannot predict how many seats you will need next year. The economics of a custom build depend on a stable population learning the same thing for several years. If that is not you, the sums do not work, and no vendor should pretend otherwise.
If two of those three describe your situation, close this tab and go and buy a library subscription. You will be better off, and so will your budget.
The test that actually separates them
Most buyers ask themselves whether their content is unique. It is the wrong question, because the answer is always yes. Everybody believes their processes are particular and their sector is unlike the others, and everybody is a little bit right, which makes the question useless as a decision tool.
Here is a better one.
What is the cost of the learner getting it wrong?
Not the cost of them disliking the course. Not the cost of a low satisfaction score. The cost of a qualified professional, holding your credential or working under your name, doing the thing incorrectly in front of a client six months later.
If the answer is "not much", buy off the shelf.
If the answer is a complaint, a claim, a regulatory finding, a disciplinary hearing, a struck-off member, a headline, or a credential that quietly stops meaning anything, then you are not really buying training. You are buying assurance, and assurance cannot be licensed from a catalogue, because the catalogue does not know what your standard is.
And one more, if you certify anything
There is a second dividing line, and it applies to a narrower group with unusual force.
Does the course carry your name?
For a professional body, a certification board, or a regulated training provider, the programme is not a support function. It is the product. It is the thing members pay for, the thing employers trust, and the thing the credential is made of.
You cannot license somebody else's course, put your crest on the certificate, and pretend the two are the same. Not because it is against the rules, but because your members will be able to tell, and the value of a credential is precisely the belief that somebody senior thought hard about what it should contain.
If the content is the product, it has to be yours. That is not a preference. It is the business model.
The money argument, honestly
You will see a lot of figures thrown around in this debate, and you should treat most of them with suspicion, because almost all of them originate in the marketing material of firms with an interest in the answer.
What can be said with reasonable confidence is this. Custom development is bought once and maintained thereafter. Licensed content is rented forever. The two curves cross, and where they cross depends on how many learners you have and how long the content stays true. With a large, stable population learning something that does not change every quarter, the custom build wins on cost eventually, and often sooner than people expect. With a small or shifting population, it does not.
But the cost conversation misses the more expensive number, which is the one nobody puts on the invoice.
An off-the-shelf course that 20% of your people complete has not saved you money. It has cost you the licence fee, plus the administrative time, plus the false assurance that your obligation is discharged, plus whatever happens later because the training did not land.
That is not a cheaper outcome than a custom build. It is a worse one, arrived at more cheaply, which is a different thing entirely.
The failure both options share
Here is the part the comparison articles leave out, and it is the only part that has ever mattered.
A custom course can fail in exactly the same way. We have seen beautifully produced bespoke programmes, built at considerable expense, with the organisation's own scenarios and its own branding and its own voice, that nobody finishes either.
Because completion is not a function of whether the content was customised. It is a function of whether anybody designed for it.
Most eLearning, custom and licensed alike, is built as a container for information and then measured on whether the information was delivered. It opens with a menu. It states its objectives. It asks the learner to remember things. And it assumes, without ever quite saying so, that a professional adult will work through twenty minutes of material out of a sense of duty.
They will not. They have a job.
The programmes that get finished are the ones designed around the learner's actual afternoon: built to earn the next click rather than to survey the subject, structured so that stopping halfway feels like leaving something unresolved, and honest about how long they take. That is a design discipline, and it is available to a custom build and unavailable to a catalogue, which is the real argument for custom eLearning and has almost nothing to do with whether the examples mention your company by name.
So the question is not custom or off the shelf. The question is whether anyone has designed for the moment your learner decides to close the tab. If the answer is no, it does not much matter which one you bought.
Where that leaves you
Buy off the shelf when the content is universal, the stakes are low, and your population is small or unstable. That is a real answer and it is right more often than my industry admits.
Build custom when the cost of getting it wrong is serious, when the credential carries your name, or when the behaviour has to change rather than merely be recorded.
And in either case, ask the harder question before you spend anything: has anybody designed this so that a busy professional will actually reach the end of it?
If you are weighing this for a specific programme
The Programme Design Diagnostic is a short, fixed-fee engagement that answers this question properly, with a written recommendation you can take to a board. It is deliberately independent of what you do next. Sometimes the recommendation is a licence.
And if you have worked through this and custom is the right call, the second question is who builds it, and how you tell a fair quote from an inflated one. That is how to choose a custom eLearning company, which is where this article hands over.
If you have decided to build, the companion pieces are how to choose a custom eLearning company and what a custom eLearning build actually looks like in 2026. See more insights from LearnFrame.