Nobody wants to use your LMS

There is a particular type of conversation I've had several times with businesses that have invested in learning technology. It normally begins with somebody showing me the platform.

This is rarely a good sign.

The platform itself is usually perfectly respectable. It has pathways, content libraries, dashboards, integrations, recommendations and increasingly some form of AI assistant that appears to know an alarming amount about everybody. Considerable effort has gone into taxonomy. Somebody has spent several afternoons deciding whether “Negotiation” belongs underneath Sales Capability or Commercial Excellence. There are reports showing exactly who has logged in, what they have completed and, if required, how long they spent completing it.

Then comes the problem.

Nobody is using it.

Or, more accurately, people are using it when the organisation makes them use it. Mandatory learning gets completed. New starters visit because onboarding sends them there. A burst of activity follows the launch of a new programme. Then usage gradually settles back towards the small population of people who genuinely enjoy browsing corporate learning platforms, who I assume also alphabetise their spice racks and read software release notes recreationally.

At this point, the conversation often becomes an adoption problem. How do we drive engagement? Should managers promote it more? Do we need internal champions? Better communications? Gamification? Perhaps another launch?

Sometimes, yes.

But before doing any of that, I think there is a more awkward question worth asking.

Why should anybody want to use it?

We confuse availability with value

Organisations have a very understandable instinct to centralise things.

Knowledge is scattered, so we create one place for knowledge. Learning is inconsistent, so we create one place for learning. Useful material lives in people's inboxes, shared drives, bookmarks, Teams channels and heads, so we purchase something with the entirely reasonable ambition of creating a single source of truth.

From an organisational perspective, this is beautifully tidy.

From the employee's perspective, it can mean we've created another place to go.

That difference matters.

Imagine I'm preparing for an important client meeting at 2pm. At 1.47, I realise I'm not entirely sure how we position a particular part of the proposition against a competitor.

I do not, at that moment, have a learning need.

I have a 1.47pm problem.

Those things may look identical to the organisation. Behaviourally, they are completely different.

If solving my problem requires me to leave the application I'm working in, remember that the information is stored in the learning platform, log in, navigate the taxonomy somebody lovingly constructed six months ago, find a 35-minute module and scrub through it until I reach the relevant three minutes, the problem is not that I lack a learning mindset.

The problem is that Google has ruined my tolerance for this sort of thing.

This is where I think businesses sometimes judge technology by the wrong standard. We ask whether the information is available when the more important question is whether accessing it is easier than whatever people will do instead.

And people always have an instead.

They message Sarah.

They search Teams.

They ask the person sitting next to them.

They use ChatGPT.

They improvise.

Or, perhaps most dangerously, they confidently do what they vaguely remember somebody telling them six months ago.

The competition for your LMS isn't another LMS.

It's the path of least resistance.

People don't adopt technology because you've bought it

There is a substantial research literature around technology adoption, and one of its more reassuring conclusions is that people aren't quite as irrational about new systems as we sometimes imply when we complain about “resistance to change”.

Theres and Strohmeier's 2024 meta-analysis brought together 134 studies of digital HR technology adoption, covering 768 effect sizes. Their analysis found that systems perceived as useful, easier to use, compatible with the user's context and supported by peers were associated with more positive attitudes and greater usage.1

Another 2023 meta-analysis, this time examining workplace technology adoption in hospitality across 30 studies and 6,728 employees, found perceived usefulness had a stronger relationship with user attitudes and intentions than ease of use.2

There is something wonderfully ordinary about that finding.

People are more likely to use technology when it is useful.

A great deal of money has been spent proving things that your grandmother could have told you.

But organisations routinely behave as though the purchase itself creates the use case. We buy the system, populate it, launch it and then become slightly offended that employees haven't incorporated it into their lives.

The behavioural question should come first.

What will somebody be doing immediately before they need this?

That is a much better starting point for learning technology than asking what content categories we need.

If the answer is “they'll be preparing for a client conversation”, support them there.

If it's “they've just been promoted and need to run their first performance conversation”, design around that moment.

If it's “they've encountered an objection they've never heard before”, giving them access to a comprehensive Negotiation Academy is technically helpful in roughly the same way that giving somebody access to the British Library is helpful when they need to know where the nearest petrol station is.

The information exists.

That's not the same as the information being useful.

Efficiency looks different from the other side of the screen

This is where I think learning technology suffers from the same problem as lots of corporate technology.

We optimise for the organisation.

One platform is easier to govern than six. Standardised content is easier to maintain than dozens of local resources. Completion data is easier to report than whether somebody made a better decision during a difficult conversation. A beautifully structured library is easier to audit than a messy network of people, prompts, coaching, search and experience.

All of that makes perfect sense from the centre.

Unfortunately, employees don't experience the organisation from the centre.

They experience Tuesday.

And Tuesday is full of meetings, clients, deadlines, messages, targets and somebody asking whether they have “five minutes” when everybody involved knows perfectly well that they do not.

Helen Tupper and Sarah Ellis make a related argument in Harvard Business Review: learning is much more likely to become part of working life when it is integrated into the flow of work rather than treated as a separate activity employees somehow need to find time for.3

Research into electronic performance-support systems points in a similar direction. Leiß, Rausch and Seifried examined how workers used digital support while solving ERP-related problems, focusing specifically on the potential for systems to support both performance and informal learning at the point of work.4

This distinction between learning technology and performance technology is useful.

The first asks:

How can we help people learn this?

The second asks:

How can we help people perform this when it matters?

Sometimes the answer to both questions is the same system.

Sometimes it absolutely isn't.

A terrible engagement metric can be a perfectly good outcome

This leads to a slightly perverse possibility.

Perhaps nobody visiting your LMS is a sign that you've designed something extremely well.

Imagine a salesperson needs help handling an unfamiliar procurement objection. They ask a question inside the tool they already use, receive a concise answer drawn from approved organisational knowledge, apply it successfully and continue with their day.

They have learned something.

They have performed better.

They have not “engaged with the learning platform” in any meaningful sense.

I would take that outcome every time.

This is where metrics can quietly distort design. Once we've invested heavily in a platform, we naturally want evidence that people are using it. Logins matter. Monthly active users matter. Content consumption matters. Completion matters.

And once those measures matter to us, we start designing things that increase them.

This is not particularly different from a restaurant deciding that the important measure of success is how many times customers look at the menu.

You can certainly optimise for it.

I'm just not sure why you would.

The useful measure depends on the problem the technology was purchased to solve.

If the problem was compliance, completion may be entirely appropriate.

If it was onboarding, perhaps we're interested in time-to-competence.

If it was inconsistent sales performance, perhaps we should see movement in the relevant sales behaviours or outcomes.

If it was managers repeatedly answering the same questions, perhaps the interesting measure is how often those interruptions happen afterwards.

If it was knowledge disappearing when experienced people leave, perhaps we need to know whether critical knowledge can now be found and applied by somebody who wasn't in the room when it was created.

Technology adoption matters because a system nobody uses cannot create much value. But usage is an intermediate outcome.

The point isn't to use the technology.

The point is to make something better.

Start with the irritating moment

McKinsey's work on digital transformation makes a similar distinction at a much larger scale. It describes transformation not as the deployment of technology but as the rewiring of how an organisation operates to create value, with technology embedded into redesigned ways of working.5 Its more recent work on AI makes the same point even more explicitly: organisations seeing greater impact are redesigning workflows around the technology rather than merely distributing tools and hoping productivity appears.6

I think there is a much simpler version of that idea which works surprisingly well when looking at learning technology.

Don't start with the platform.

Start with the irritating moment.

• Where do people repeatedly get stuck?
• What question do managers answer for the tenth time that week?
• Where does a salesperson know that the answer exists somewhere but can't find it quickly enough to be useful?
• What do new starters continually misunderstand?
• Where does somebody have to stop doing productive work in order to find out how to do productive work?
• What important judgement currently depends on remembering which experienced person knows the answer?

Find those moments first.

Then decide what the best response is.

Perhaps it is an LMS. Perhaps it's an LXP. Perhaps it's an AI layer over organisational knowledge. Perhaps it's a prompt inside the CRM, a searchable job aid, a coaching conversation, a redesigned process or a person.

The technology is interesting only after the problem is interesting.

I worked with one organisation where the initial conversation was essentially about making better use of the learning platform they already owned. Usage was lower than expected, and the assumption was that the organisation needed to drive people towards it more effectively.

Once we looked at the situations in which people actually needed support, the question changed. The problem wasn't primarily that employees were unwilling to learn. They were already learning constantly, usually by asking colleagues, searching old conversations or improvising from experience. The platform simply wasn't where the problem occurred, so using it required an additional behavioural step that offered them very little immediate reward.

That is an expensive thing to solve with communications.

Sometimes low adoption is a change-management problem.

Sometimes it is a design problem.

And sometimes it is the workforce giving you perfectly rational feedback about a system nobody asked them whether they needed.

Before spending another pound increasing engagement with your LMS or LXP, I'd therefore do something slightly counterintuitive.

Stop looking at the platform analytics for a week.

Spend that week looking at the work.

Find five moments where somebody needs knowledge, guidance or capability in order to perform something important. Watch what they actually do. Notice where they look, who they ask, what frustrates them and how quickly they need an answer.

Then ask whether your technology makes that behaviour easier.

If it does, adoption becomes a considerably simpler problem.

If it doesn't, another internal launch campaign may produce a pleasing spike in logins.

But I wouldn't mistake that for value.

RB