Why measurement is hard for L&D
If you've ever been asked "what's the ROI of training" and felt a wave of dread, you're in good company. L&D, internal comms, and enablement teams sit at the intersection of three measurement problems:
The work is upstream. Training doesn't generate revenue directly. It enables other people to generate revenue, or reduces the cost of bad outcomes that didn't happen. Both are real, both are hard to attribute cleanly.
The audience is diverse. A single training programme might serve sales, support, ops, and engineering. Each group cares about different outcomes, on different timelines, measured in different systems.
The signal is delayed. A new hire trained today produces value in 90 days. A compliance refresher pays off only when an incident doesn't happen. Most measurement frameworks built for marketing or sales assume a much tighter feedback loop.
None of this means measurement is impossible. It means starting with the right framework matters more than picking the right tool.
The three-tier framework
Most L&D teams measure something. Few measure the right things at the right level. A useful way to think about it: every metric you track sits in one of three tiers, and a healthy measurement programme has at least one metric in each.
Tier 1: Business outcomes. What the CFO sees. Time to productivity. Retention. Performance lift. Compliance pass rate. These are the metrics your work ultimately moves, even if the attribution is partial.
Tier 2: Programme effectiveness. How well your programmes are landing with learners. Completion. Engagement. Knowledge retention. Time to update content when something changes. These are the metrics you use to manage day-to-day.
Tier 3: Production efficiency. How efficiently you're producing the work itself. Content volume. Cost per asset. SME time required. These are operational metrics for capacity planning and budget conversations.
The mistake most teams make is living entirely in tier 3 (output metrics) or tier 2 (completion rates) and never connecting up to tier 1. The result is an L&D function that can describe what it produced but not why it mattered. The fix isn't to abandon tier 2 and 3 - it's to anchor a small number of tier 1 metrics that tier 2 and 3 ladder up to.
Tier 1: Business outcome metrics
The metrics that tie L&D work to business results. If you can move these, you're a strategic function. Pick one or two to anchor on.
- Time to productivity / time to competency
- What it measures: How long it takes a new hire or role-changer to reach full performance.
- Why it matters: Directly tied to revenue, customer experience, and operational efficiency. A 10% reduction here is a board-level number. Particularly powerful in sales, support, and operations roles where productivity is measured.
- Where to find it: HR usually tracks ramp time for revenue roles. For others, ask hiring managers for their estimate of "fully productive" - the answer often varies widely and the conversation itself is valuable.
- Common targets: 10–20% reduction over 12 months is meaningful. Bigger reductions are possible but usually require process changes beyond training.
- Watch for: Definitions of "productive" vary by function. Agree the definition before you start measuring.
- Employee retention
- What it measures: Percentage of employees still in role at a given interval (commonly 90 days and 12 months), often broken down by cohort or function.
- Why it matters: Onboarding quality is one of the top predictors of early retention. If L&D owns the onboarding experience, you own a meaningful slice of the retention story.
- Where to find it: HR systems. Often segmented by manager, function, or office.
- Common targets: 1–3 percentage point improvement over a year is genuinely meaningful, especially in high-volume hiring or roles with historic churn.
- Watch for: Many things drive retention. L&D's contribution is real but partial. Frame as "supporting evidence" not "we caused this."
- Performance lift
- What it measures: Measurable improvement in business KPIs that can be partially attributed to training. Sales productivity, call handle time, defect rate, NPS, claim accuracy, conversion rate - whatever the function being trained is measured on.
- Why it matters: The cleanest way to prove L&D contributes to business outcomes. Particularly powerful if you can run a before/after comparison on a specific cohort.
- Where to find it: Whatever system the function uses (CRM, support platform, ops dashboards). Partner with the function lead to get the right number.
- Common targets: Even a 2–3% lift on a large team is material. Don't promise more than you can defend.
- Watch for: Attribution is genuinely hard. The most credible version is "this cohort received the new training, this comparable cohort didn't, here's the difference." Without a comparison group, treat results as directional.
- Compliance pass rate / incident reduction
- What it measures: Percentage of employees completing required compliance training within the regulatory window, and the rate of compliance incidents over time.
- Why it matters: In regulated industries (financial services, healthcare, manufacturing, logistics, utilities), compliance failures carry real financial and reputational cost. Compliance L&D is often the most senior, best-funded part of the function for a reason.
- Where to find it: Compliance team, audit reports. Often already tracked with high rigour.
- Common targets: 95%+ on-time completion is the floor in most regulated industries. The harder metric is incident rate, where even small reductions translate to material risk avoidance.
- Watch for: Don't conflate completion with comprehension. Someone clicking through a course doesn't mean they'd handle the situation correctly. Pair with knowledge checks where possible.
Tier 2: Programme effectiveness metrics
The operational metrics for managing programmes day-to-day. These tell you whether your content is landing, even before you can see the business outcome it drives.
- Completion rate
- What it measures: Percentage of assigned learners who finish a course, module, or programme.
- Why it matters: Historically the headline L&D metric. Still used everywhere because it's easy to measure. But on its own it tells you very little, a 95% completion rate on a bad course is worse than a 60% rate on one that actually changes behaviour.
- Where to find it: LMS or learning platform.
- Common targets: Floor of 80%+, peaks above 95% for mandatory content. Be more interested in the shape of the dropout curve than the headline number.
- Watch for: Easy to game. Don't make completion the primary KPI. Use it as supporting evidence behind a tier 1 metric.
- Learner engagement and satisfaction
- What it measures: CSAT, NPS, qualitative feedback, time-in-course, voluntary re-watching or revisiting.
- Why it matters: A signal of content quality and learner trust. Low engagement is often the precursor to low completion and low impact. High engagement doesn't guarantee impact, but low engagement almost always undermines it.
- Where to find it: Post-course surveys, learning platform analytics, qualitative interviews with a sample of learners.
- Common targets: Half-point to one-point improvements on a 1–5 scale are meaningful when the population is large. Track trends over time more than absolute scores.
- Watch for: Survey fatigue is real. Keep feedback mechanisms short and focused.
- Knowledge retention
- What it measures: How much of the content learners remember and can apply, typically measured at 30, 60, or 90 days post-training.
- Why it matters: Completion without retention is wasted spend. If knowledge fades within a month, the programme isn't doing its job regardless of completion rates.
- Where to find it: Follow-up quizzes, manager observation, sample-based interviews.
- Common targets: 10–20% improvement vs a baseline (or vs no-training control) is meaningful.
- Watch for: Hard to measure rigorously without a measurement framework already in place. Start small — one programme, one cohort, before scaling.
- Time to deploy / time to update
- What it measures: How long it takes to get new content into learners' hands, and how long to update existing content when something changes.
- Why it matters: A leading indicator of L&D agility. Slow update cycles mean stale content, which kills credibility and engagement. In regulated industries or fast-moving product environments, this is often the most strategic metric you can move.
- Where to find it: Project management tools, intake-to-launch tracking, or simple manual measurement on a few representative pieces.
- Common targets: Varies wildly by content type. The right question is "is this faster than last quarter" rather than hitting an industry benchmark.
- Watch for: Often unmeasured because no one's asked. Start measuring it and you'll likely find it's worse than you thought, which is the point.
Tier 3: Production efficiency metrics
Day-to-day operational metrics for capacity planning, budget conversations, and process improvement. Useful as supporting evidence rather than headline metrics.
- Content production volume
- What it measures: Number of pieces of content produced per period, often broken down by content type (course, microlearning, video, comms).
- Why it matters: Capacity planning. Demonstrates the function is keeping up with demand from the business.
- Where to find it: Internal content tracking. Often already exists informally.
- Common targets: Year-on-year growth, or volume against a planned capacity.
- Watch for: Volume without quality is meaningless. Always pair with a quality or engagement metric.
- Cost per learner or cost per content asset
- What it measures: Total L&D spend divided by learners served, or by content produced. Used for budget conversations and benchmarking.
- Why it matters: Defending budget. Justifying new investment. Identifying programmes that are unusually expensive relative to outcomes.
- Where to find it: Finance partner. Often requires some allocation logic.
- Common targets: Reduction over time as scale increases. Significant variance between programmes is normal and worth investigating.
- Watch for: Cheaper isn't always better. A higher cost per learner on a programme with measurable business impact is fine. The metric is most useful for comparison, not as an absolute target.
- SME time spent in production
- What it measures: Hours of subject matter expert time required to produce a piece of content.
- Why it matters: SME time is usually the scarcest resource in L&D production. Reducing it unlocks capacity for everything else and removes a common bottleneck.
- Where to find it: Ask SMEs. Often shockingly high once surfaced.
- Common targets: 50–80% reduction is achievable with the right tools and workflows.
- Watch for: SMEs often underestimate their time investment because much of it happens in fragments. Track it carefully for one or two programmes to get a realistic baseline.
- Localisation coverage and cycle time
- What it measures: Number of languages content is available in, and time to make new content available in additional languages.
- Why it matters: Global teams are usually under-resourced for the languages they need to cover. Often an unmet need rather than a tracked KPI, which is itself a problem.
- Where to find it: Self-audit. Most teams discover their coverage is narrower than assumed.
- Common targets: Coverage growth in line with workforce or customer distribution. Cycle time from weeks to days.
- Watch for: Translation quality matters as much as quantity. Don't ship content in 20 languages if half of them are unreviewed.
Where to start if you're not measuring anything
A practical sequence for teams building a measurement programme from scratch.
Pick one tier 1 metric. Just one. The one most relevant to your function's biggest perceived problem. If onboarding is the pain point, time to productivity. If you're in a regulated industry, compliance pass rate. If you sit close to revenue, performance lift on a sales or support metric. Don't try to measure everything.
Find the baseline. Whatever you pick, find the current number. Talk to HR, Finance, or the function lead. If no one's tracking it, that's a finding in itself — start tracking it now.
Add one tier 2 metric that supports it. Time to deploy or completion or satisfaction, depending on what feels most relevant. This is your operational health check.
Set a 6–12 month target. Be realistic. Improvement of 10–20% on a tier 1 metric is meaningful and defensible. Bigger numbers invite scepticism.
Review quarterly. Don't wait for the annual review. Measurement that doesn't drive conversations doesn't drive change.
Add more over time. Once one programme is well-measured, add a second. Resist the urge to measure everything at once as it dilutes attention and rarely improves outcomes.
Common pitfalls
- Measuring only completion
- The most common L&D mistake. Completion is easy to measure but disconnected from impact. A function that lives in completion rates struggles to prove value when budgets tighten. Always pair it with a tier 1 metric.
- Trying to measure everything
- Teams new to measurement often try to track 15 metrics across every programme. This sounds rigorous but produces noise rather than signal. Pick a few, measure them well, add more over time.
- Picking metrics you can't influence
- It's tempting to anchor on a big strategic number like "employee NPS" because it sounds important. If L&D's contribution to that number is small and noisy, you'll struggle to defend it. Pick metrics where your work makes a visible difference.
- Claiming pure attribution
- Almost no L&D outcome is purely attributable to training alone. Better managers, better tools, better processes all play a role. Frame your contribution as supporting evidence ("the cohort that received the new training showed X improvement") rather than sole cause. It's more honest and more defensible.
- Treating qualitative as not-real
- Quotes, interviews, and observations from learners and managers are data. They're not statistically rigorous, but they often surface things metrics miss. Use them alongside numbers, not instead of them.
- Not reviewing what you measure
- A metric defined once and never revisited becomes wallpaper. Build a quarterly cadence for reviewing whether the metrics are still the right ones, whether you're hitting them, and what to do if you're not.
Beyond the metric: how to make measurement land internally
A measurement programme that produces beautiful dashboards no one looks at is worse than no programme at all. A few principles for making the numbers matter inside the organisation.
Find a partner outside L&D. The strongest L&D measurement stories are co-owned with the function being trained. The sales enablement metric belongs to L&D and the CRO. The onboarding metric belongs to L&D and HR. Joint ownership creates joint accountability and makes the story land harder upward.
Tell the story, don't just show the number. A 12% improvement in time to productivity is interesting. The same number paired with "which means each new hire generates an extra £40K of revenue in their first 90 days, and we onboard 200 people a year" is a budget conversation. Always convert metrics into the language your audience speaks.
Make it easy for your stakeholders to use the numbers. If your CFO can pull a slide on L&D impact and drop it into a board pack without rewriting, you've won. If they have to re-interpret your data, you've made their life harder and they'll stop asking.
Be honest about what you can't measure. Trust is built by being transparent about the limits of attribution, not by overclaiming. A function that says "we contributed to this, alongside these other factors" is more credible than one that claims sole credit for everything.
Revisit when the business changes. Reorgs, new strategies, leadership changes all shift what matters. Metrics that mattered last year might not be the right ones now. Refresh annually.
This guide is provided as a starting point for L&D, comms, and enablement leaders working through measurement questions. None of the frameworks here are specific to any tool or platform — they're industry conventions adapted for teams that need to demonstrate value internally. If you'd find it useful to walk through how to apply this to your specific context, the person who shared this with you can help.
