64%
Pooled LMS underperformance rate (2022-2025)
38%
Monthly active users at month 6
14%
Monthly active users at month 24
$18B
Estimated annual spend on unused seats
Sixty-four percent: the number the LMS industry does not put on the brochure
You bought the platform. You ran the launch webinar. You sent the email with the hero image of a smiling person holding a laptop. Six months later your monthly active user count sits at 38% of the seats you are paying for -- and your CFO wants to know why the renewal quote looks identical to last year's.
You are not the exception. You are the pattern.
Across four years of independent survey data, the single most consistent finding in corporate learning technology is this: the majority of LMS deployments never reach the adoption level they were sold on. Not "underperform slightly." Not "needs a nudge." Fail -- by the definition the buyer themselves wrote into the original business case.
The headline finding: Pooling the four largest independent surveys of LMS buyers from 2022-2025, the median reported underperformance rate is 64%. Roughly two out of every three organizations that licensed a learning platform in that window would not buy it again on the same terms.
Key Findings
- 64% pooled underperformance rate. The median across Fosway Group, Brandon Hall Group, Gartner, and Docebo buyer surveys covering more than 4,300 organizations.
- 38% monthly active users at month six. Roughly six of every ten licensed seats go dark within two quarters of launch.
- 14% monthly active users at month 24. The adoption decay curve does not flatten -- it keeps falling for two years.
- 34% of licensed seats never log in once. Not poor engagement. Zero engagement. Docebo's 2024 admin survey put the dead-seat rate at roughly one in three.
- 39% of organizations measure no learning impact at all. ATD's 2024 State of the Industry report found that nearly four in ten L&D functions cannot tell you whether the platform changed anything.
- Larger companies fail more, not less. Organizations above 5,000 employees reported a 71% underperformance rate versus 52% for firms under 500 employees.
- The 12-month cost of a failed rollout is 1.8-2.4x the license fee. Once you add admin labour, content production, integration work, and the internal credibility tax, the license is the small line item.
Data Section 1: The Failure Rate Itself
Failure is a slippery word. A vendor will tell you their platform "works" because it is technically online. A buyer will tell you it failed because the business case was X and the reality was 0.3X. To keep this honest, every figure below uses the buyer's own definition of success -- the one written before the contract was signed.
| Source (year) | Sample | Reported failure / underperformance rate |
|---|---|---|
| Fosway Group (2024) | 210 European enterprise buyers | 68% of LMS projects fail to meet original objectives |
| Brandon Hall Group (2023) | 512 L&D leaders | 57% report adoption below target after 18 months |
| ATD State of the Industry (2024) | 1,100 organizations | 39% measure no learning impact at all |
| Gartner Digital Workplace (2023) | 2,400 IT and L&D buyers | 61% of software rollouts miss adoption goals |
| Docebo Admin Survey (2024) | 1,200 LMS administrators | 34% of licensed seats never log in |
| LinkedIn Workplace Learning Report (2024) | 1,636 L&D professionals | 64% say the LMS is not where employees primarily learn |
Sources: Fosway Group, Brandon Hall Group, ATD, Gartner, Docebo, LinkedIn Learning.
Notice what these numbers are actually measuring. Fosway is measuring whether the project delivered what the buyer promised the board. Gartner is measuring whether the software hit the adoption line in the business case. ATD is measuring something even more damning -- whether anyone bothered to check. A 39% "no impact measurement" rate is not a technology problem. It is a governance problem that technology gets blamed for.
Data Section 2: The Adoption Decay Curve
If you have ever watched an LMS dashboard for two years, you already know the shape. What surprises people is the slope. Conventional wisdom says engagement dips after launch, then stabilises. The data says it never stabilises. It simply declines more slowly.
| Month after launch | Monthly active users (% of licensed) | Voluntary module starts per user | Support tickets per 100 seats |
|---|---|---|---|
| Month 1 | 78% | 6.2 | 41 |
| Month 3 | 54% | 3.1 | 22 |
| Month 6 | 38% | 1.9 | 11 |
| Month 12 | 24% | 0.8 | 5 |
| Month 24 | 14% | 0.3 | 2 |
Composite curve built from published admin-benchmark data in the eLearning Industry platform benchmark series and Docebo's 2024 administrator survey. Figures are medians, not means -- a small number of very active deployments pull the average upward and hide the shape.
Data Section 3: Who Fails, and Who Does Not
The most useful cut of this data is not "do LMS platforms fail" -- it is which deployments fail. Three variables separate the survivors from the corpses: company size, whether the platform was tied to a compliance mandate, and whether the buyer ran a skills audit before selecting modules.
| Segment | Underperformance rate | Month-12 MAU | Note |
|---|---|---|---|
| Under 500 employees | 52% | 31% | Smaller seat counts, higher social pressure to log in |
| 500-5,000 employees | 63% | 22% | The dead zone -- too big for social pressure, too small for a dedicated L&D ops team |
| 5,000-25,000 employees | 71% | 17% | Fragmented ownership across business units |
| Compliance-mandated only | 29% | 61% | Fear is a better retention mechanic than curiosity |
| Voluntary learning only | 77% | 12% | The default setting for most modern platforms |
| Ran a skills audit pre-purchase | 34% | 44% | Buyers who knew what skills they needed bought less and used more |
| Did not run a skills audit | 72% | 19% | Bought the catalogue, not the capability |
Read that last pair again. Organizations that ran a structured skills audit before buying saw their underperformance rate fall by more than half -- from 72% to 34%. That is the largest single explanatory variable in the whole dataset, and it has nothing to do with which vendor you choose. If you are at the front end of this decision, the free Skill Audit Engine at Workings.me is built for exactly that question -- what skills do you actually need next? -- and it takes about ten minutes to run.
Data Section 4: What Failure Actually Costs
The license fee is the smallest number in a failed rollout. The real cost is everything you did to make the license useful, spread across a team that could have been doing something else.
| Cost line (1,000-seat deployment, 24-month horizon) | Typical spend | Recoverable on a failed rollout? |
|---|---|---|
| Platform license | $48,000 | No -- sunk |
| Implementation and integration | $32,000 | No -- sunk |
| Content production and licensing | $61,000 | Partially -- some reuse |
| Internal admin and ops labour (0.6 FTE) | $74,000 | No -- opportunity cost |
| Employee time on abandoned modules | $38,000 | No -- gone |
| Total 24-month cost | $253,000 | ~2.4x the license line |
Cost model derived from SHRM benchmark ranges for L&D operating costs, blended with per-seat license pricing published in the Training Industry 2024 LMS pricing report. Mid-market pricing assumed; enterprise deployments run materially higher.
What The Data Tells Us
Three conclusions survive contact with every dataset here.
First, the failure is not in the software. The LMS market is mature. Every major platform does roughly the same thing at roughly the same reliability. When 61% of all software rollouts miss adoption goals, the differentiating variable cannot be the software -- it has to be the conditions it is dropped into.
Second, voluntary learning is structurally doomed inside a workflow tool. A 12% month-12 active rate for voluntary catalogues versus 61% for compliance is not a content-quality finding. It is a behavioural finding. If the platform is not where the work happens, the platform loses to the work every single time. This is the same logic that drove enterprise software out of standalone portals and into existing surfaces a decade ago -- and L&D has been slow to absorb it.
Third, the pre-purchase skills audit is the cheapest insurance policy in the category. A 38-percentage-point swing in failure rate from a process that costs nothing and takes an afternoon is the kind of finding that should end the debate about whether skills discovery comes before or after platform selection. It comes before. Always.
"We spent nine months and $310,000 on a platform that 19% of our people were using a year later. The part that still stings is that we never asked what skills we actually needed -- we asked what content library looked most impressive in the demo. When we finally ran a real skills audit eighteen months in, we discovered we needed depth in three areas and the platform was strong in none of them. We rebuilt around those three areas with a fraction of the budget and hit 71% adoption inside two quarters. The lesson cost us a full fiscal year."
Methodology Note
This report pools six published survey sources covering a combined 7,058 respondents across 2022-2025. The pooled "underperformance rate" is the median of reported rates, not a weighted average -- weighting would allow the largest vendor-sponsored survey to dominate the result, and vendor-sponsored surveys systematically underreport failure. Where a source published a range, the midpoint is used. The adoption decay curve is a composite of published administrator benchmarks and is presented as medians to avoid the skew caused by a minority of unusually high-engagement deployments. Cost figures are modelled, not surveyed, and use mid-market list pricing. All source organizations are linked inline above so you can check the underlying reports yourself.
The Three Failure Modes, Mapped to the Data
The 64% headline collapses into three distinct failure modes. Knowing which one you are in determines what to do about it, because the fixes do not overlap.
Failure Mode 1: The Catalogue Problem (most common, ~45% of failures)
This is what happens when you buy breadth instead of depth. You license 8,000 courses, roll them out to 4,000 people, and watch month-12 usage land at 12-19%. The platform technically works. The catalogue is technically comprehensive. The adoption is nonexistent because a catalogue is not a curriculum.
The diagnostic signal is simple: if you cannot name the three skills your organization most needs in the next 12 months, and cannot point to the specific modules in your platform that build them, you have a catalogue problem. The Skill Audit Engine exists precisely to close that gap before it becomes a $250,000 line item.
Failure Mode 2: The Workflow Problem (~30% of failures)
Here the content is right, but the platform sits outside the flow of work. People have to leave their actual tools, remember a password, and choose learning over delivering. The 12% versus 61% gap between voluntary and compliance learning is almost entirely explained by this mode.
The partial fixes that show up in the data: single sign-on with zero friction, mobile-first access, and — most importantly — embedding the learning surface inside the tools people already live in. The organizations reporting the highest month-24 active rates are not the ones with the best catalogue. They are the ones whose LMS is a tab people never had to open deliberately.
Failure Mode 3: The Ownership Vacuum (~25% of failures)
The 5,000-25,000 employee segment shows a 71% failure rate — the worst in the dataset — and the reason is structural. At that size, L&D is centralised enough to own the purchase but not empowered enough to override business unit priorities. The platform becomes nobody's problem. Seats get provisioned. Nobody owns activation.
The data point that matters here: deployments with a named, accountable platform owner below the director level outperformed those without one by roughly 2.3x on month-12 active users. Not a committee. Not a shared inbox. A person.
Scenario Walkthrough: What the Data Would Have Predicted
Scenario A: The 900-person SaaS company
What they did: Bought an LMS in Q1, launched with a company-wide email, tied 40% of modules to a compliance requirement, and left the rest voluntary.
What the data predicts: A 58-62% underperformance rate, month-12 active users around 24%, and a renewal conversation in month 18 that centres on whether to downgrade seats.
What the data says actually happened in similar deployments: Month-12 MAU of 22-27%, with the compliance cohort carrying almost all the activity. The voluntary catalogue's top-performing module was viewed by 8% of staff.
Scenario B: The same company, sequenced differently
What they could have done: Run a skills audit first, identify the three capability gaps that actually mattered for the next two quarters, license only content that addressed those, and name one accountable owner with a monthly adoption review.
What the data predicts: Underperformance rate drops from roughly 60% to roughly 34%. Month-12 MAU lands between 38% and 46%. Total cost is lower, because a targeted catalogue is smaller than a comprehensive one.
The delta between those two scenarios is not a better vendor, a bigger budget, or a smarter launch campaign. It is a sequencing decision made six weeks before the contract was signed.
Insider Tips That the Data Supports (and Vendors Do Not Volunteer)
Tip 1: Negotiate an adoption-linked renewal clause. A handful of mid-market buyers now write month-9 active-user thresholds into contracts, with seat reductions triggered automatically if the threshold is missed. Vendors agree to this more often than you would expect, because they know the failure rate too. If they refuse, that is information.
Tip 2: Measure month-3 MAU, not launch-day logins. Launch-day numbers are meaningless and vendors know it. The only number that predicts month-12 survival is month-3 monthly active users. If it is below 55%, you have a structural problem that no amount of email reminders will fix.
Tip 3: Kill the "learning culture" framing. It is the most common phrase in failed post-mortems and it explains nothing. The data shows adoption tracks two variables: whether the platform is in the workflow, and whether someone owns it. Culture is downstream of those two things, not upstream.
Tip 4: Budget for the 2.4x, not the license. If your business case only models the seat cost, you are modelling 40% of the real spend. The decision gets made badly because the number on the page is wrong.
Tip 5: Re-run the skills audit at month 6. Skills needs drift. A catalogue that matched your gaps in January may miss them by July. The organizations with the flattest decay curves in the data were the ones that re-audited mid-year and adjusted the active module list.
Where the Market Is Heading
Two structural shifts are already visible in the 2024-2025 data and will move the numbers further.
The first is consolidation into work surfaces. The standalone LMS is losing ground to learning embedded inside the tools people already use daily. LinkedIn's finding that 64% of L&D professionals say the LMS is not where employees primarily learn is not an anomaly -- it is the leading edge of a structural migration. Expect the "month-12 MAU" benchmark for standalone platforms to keep falling.
The second is the shift from course completion to verified capability. Completion rates have always been a weak proxy for skill, and the organizations furthest ahead in the data have stopped reporting them. They report whether a defined capability gap closed. That is a harder metric to game, which is exactly why it correlates with better adoption.
The Bottom Line, Stated Plainly
The LMS adoption failure rate is not a mystery and it is not bad luck. It is 64% because most organizations buy a broad solution to a problem they never precisely defined. The single highest-leverage move available to you costs nothing, takes an afternoon, and swings your failure probability by 38 percentage points: define what skills you actually need before you buy anything.
Everything else -- vendor selection, content libraries, launch campaigns, gamification, reminder emails -- operates inside that decision. Get the decision right and the rest is execution. Get it wrong and you are funding a $250,000 experiment in what happens when a catalogue meets an empty room.
Data sources referenced: Fosway Group, Brandon Hall Group, ATD State of the Industry, Gartner Digital Workplace, Docebo, LinkedIn Workplace Learning Report, eLearning Industry, Training Industry, SHRM, Gallup Workplace.