Proving L&D Impact: Measure the Effect, Not the Activity

Proving L&D Impact - Measure the Effect, Not the Activity

L&D teams have more data than ever, but how much of it actually proves anything? L&D experts Lee Cooper and Tom McDowall unpack how to measure the true impact of your work.

The plan for most reporting isn’t for it to be a list of activities.

But it happens over time.

You share completions here, attendance there, a happy sheet to round it off.

On paper, it looks impressive. 5,000 people trained. 90% completion. 

Everyone enjoyed the lunch.

But ask what changed in the business, and the answer often isn’t there.

No other team reports like this. 

Your sales team don’t tell the board how many calls they made. 

They tell them what they sold.

And with only 12.7% of learning and development professionals confidently reporting ROI, teams now more than ever need to prove the difference they make. 

Not how busy they are.

Lee Cooper joined Tom McDowall for the third episode of BuildEmpire’s webinar series in partnership with Totara in November, split into what L&D impact really means and how to actually prove it. 

Their throughline: completions show you were busy. Impact shows you made a difference.

⏰ TL;DR

Start with the effect the business wants, then plan backwards and report forwards, with participation and ability data as supporting evidence rather than the headline. Get managers and data owners involved early, because the change you’re trying to prove happens on the job, not in the course. And communicate it as one headline in business language with a clear recommendation, because data isn’t the finish line; decisions are.

Keep reading to learn:

  • What is L&D impact?
  • Why completion data still matters
  • Why managers hold the key to proving impact
  • How to make sense of all the data
  • How to communicate L&D impact to the business
  • How LMS data can reveal accessibility barriers
  • Where to start with proving L&D impact

What is L&D impact?

L&D impact is the measurable difference learning makes to people’s performance and to the organisation.

It’s not how many courses you’ve built, how many people completed them or how highly they rated the session.

Those are activity measures. They tell you something happened.

Impact tells you what changed because of it: people doing their jobs differently, and the business seeing the result, whether that’s lower attrition, fewer incidents, faster time to productivity or more sales.

It can look different in every organisation. And that’s the point. The impact that matters is the one the business has agreed it wants to see.

Lee Cooper, Senior Solutions Manager at Totara, has spent over 25 years in L&D and HR, after time in the armed forces and retail finance. 

He got into learning technology when he was asked to train 4,000 people across the country on anti-money laundering procedures, and couldn’t see how to do it alone.

Like most people in L&D, he got into it to make a difference. 

To explain what that looks like, he uses a framework called PACE:

  • Participation: what did people do?
  • Ability: are they capable of doing it?
  • Change: are they doing something different?
  • Effect: did the business improve?

The problem is where most reporting stops. “Most of the reporting that we tend to do in learning technologies or the organisations I work with stops at P.”

Did people complete the course?

Did they attend the seminar?

Proving impact means getting from P all the way to E. 

As Lee put it, it’s the difference between saying you’ve created a thousand courses and asking, “what was the impact of those thousand courses?”

Here’s how to make that shift, and what it should change about how you plan, measure and report learning.

Why completion data still matters

It’s become a bit trendy to dismiss completion data altogether. It’s useless; it doesn’t tell you anything, don’t even look at it.

Tom, a user of Will Thalheimer’s LTEM model, thinks that’s unfair.

Participation data is the easiest to get, which is exactly why L&D has so often stopped there. 

As Lee put it, “we’ve made a rod for our own backs” by jumping on completions.

But the data itself isn’t the problem. How you use it is.

Lee’s approach is to “plan backwards and report forward.”

Start with the effect the business wants to see. 

Then tell the story from the beginning, with participation as the opening chapter, not the ending.

Say 90% of new managers completed a programme. On its own, that’s a number. The useful question is what’s happened since.

Tom flipped it the other way too. Why did the other 10% not take part?

Is it a technical problem? The wrong modality or location? Or “a lack of faith in the internal L&D team, which is really important for us to know about”?

Participation data tells you whether you’re succeeding at a basic level. 

Can people access the course?

Is it in the right place?

Is something stopping them from finishing?

Ability data works the same way. A quiz score on its own doesn’t mean much. 

But put it next to real-world performance, and you’ve got two data points worth comparing.

Do that often enough, Tom said, and you can start building a basic predictive model: people who score 75% or 80% plus can reasonably be expected to perform within a certain range.

“That suddenly sounds much more like what the finance department might do or the sales team or the legal team,” where data drives decisions, rather than saying, “we trained 5,000 people. I mean, they all quit, but we trained them.”

Related: Complete guide to LMS analytics

Why managers hold the key to proving impact

Here’s the awkward bit. L&D rarely owns the data that proves impact.

It usually sits with finance, operations or sales. 

And there’s often a reluctance to go beyond the data L&D and HR already have.

Lee’s view? The business would probably welcome it.

For years, L&D asked for a seat at the table. 

Now it’s got one, and with business partners embedded across finance, operations and the shop floor, it has access it never had before. “We shouldn’t kind of ignore that.”

The biggest factor of all is the manager.

Managers do the day-to-day coaching. They’re the ones who’ll notice if somebody’s doing something different, or if their behaviour’s changed.

In PACE terms, the route from A to C, from can they do it to are they doing it, runs through the manager. 

“The manager is normally the person that has the answers.”

And it’s very rarely more content. “We don’t need, you know, a top-up course.”

Tom added a distinction that’s more than semantics: learning impact versus performance impact.

“We are not the workplace education team. We are not looking for a highly learned workforce.” The goal is a performant workforce, whatever performance means in your organisation.

Lee put it simply. Learning impact proves the training worked. Performance impact proves the organisation benefited. “The difference between these two is what happens after the course, and that’s usually where the manager comes in.”

That matters because most performance problems have more to do with the environment than with individual knowledge.

Managers control a lot of that environment: time, pressure, support, motivation. L&D can make sure people are able to do something. It can’t take those final steps with them.

How to make sense of all the data

Manager observations, HRIS data, attrition rates, time to productivity, 360 feedback, business KPIs. That’s a lot coming in.

Tom asked how you put some order to that chaos. Lee’s honest answer: “With difficulty.”

2026 analysis from D2L reported that 95% of L&D organisations don’t excel at using data to align learning programs with business objectives, so here are a few things to make it manageable.

Start with a baseline. You need to know what was happening before, so you can see what’s changed.

Add comparison groups, like trained people versus not-yet-trained people. “It’s almost like a science experiment, isn’t it?”

Then look for trends over time. Baseline, comparison, trend: “if you have those three… then you have a credible story.”

And look at the same story from several sources. A manager rating, a business metric and a learner’s own account together are far stronger than any one of them in isolation.

You’re rarely going to prove causation, Lee admitted. 

So aim for the evidence the business will believe. Use their own numbers. Trust finance’s figures. Don’t invent new metrics.

Success cases help too. The success case method finds real examples that bring the numbers to life.

On AI, Lee was careful. He’d never advocate putting personal data into ChatGPT. 

But beyond that, AI can help L&D spot trends it might otherwise miss. 

Tom’s bet is on “small AI”: internal models your organisation owns and knows are safe.

Both agreed on one thing: collect more than you think you need.

Too often, teams decide on three numbers up front and ignore the rest. 

Tom’s stance is to measure everything. “You’re interested in everything. You just don’t know it yet.”

Interfere with something as complicated as a workplace and unintended things happen. Some good, some bad. 

Revenue might go up, but costs might come down too, and if you’re not measuring it, you miss the chance to say so.

Lee’s version came from his football card collection. He’s surrounded by thousands of them and wishes he’d started categorising them from day one, “because now the job is too big.” 

Don’t make your future self unhappy.

Tom’s favourite example was a biotech client that tracked a value calculation for every employee: what it cost to hire them, everything invested in them since, and the revenue their work generated.

When someone left, they knew exactly what that cost. 

Imagine designing an induction programme with that insight, and asking what was different about how the highest-value people were onboarded.

That’s the bigger point. Impact data isn’t just for proving value afterwards. It should inform your design decisions and priorities too.

And when the data looks perfect, be suspicious. “If the data is telling you exactly what you wanna hear, don’t trust it.”

A spreadsheet full of green cells is good, but very boring. It’s the solitary amber or red cell that holds the interesting question: what’s holding that back?

How to communicate L&D impact to the business

You’ve got the numbers. You think you understand them. Now someone else has to.

And chances are, your C-suite sponsor or line managers don’t have time to trawl through a spreadsheet with a couple of graphs attached.

Lee always starts by going back to the original question: what effect did the business agree it wanted?

With so much data coming in, it’s easy to get bogged down and end up proving something different from what the business asked for.

Then he turns the whole analysis into one headline. A single “so what?”

For example: teams whose managers completed the programme had lower early attrition.

That’s it. That’s the story.

From there, tell it in the business’s language. Lead with the effect, then work back through change, ability and participation as the supporting evidence.

Give a one-page story, not a dashboard full of charts.

Tom learned that the hard way. Early in his career, he sent a 30-page deep-dive to a C-suite team, with the recommendations on page 29. 

The response? They weren’t reading it. If you can’t fit it on one page, you don’t know what you’re talking about.

People are messy, and the stories behind impact are complicated. But “it’s complicated” isn’t a report. Here’s what we think, here’s what we know, here’s what we need to do next.

Tom’s advice is to go one step further than helping people decide. Tell them the decision you think they should make.

That’s how L&D stops being an order taker. “Is it okay if we continue to improve the business? That’s a much easier yes than would you like to spend some money on some more training, boss?”

And the answer won’t always be more learning. Lee pointed out the evidence might show the fix is something else entirely, and that’s just as valuable.

Whatever it shows, act on it. Scale what worked. Fix or stop what didn’t. 

Brief managers. Make reporting routine, not a one-off.

“Data is not the finish line. Decisions are.”

Or as Lee put it, if your report doesn’t change what anyone does next, it’s still an activity report.

How LMS data can reveal accessibility barriers

One of the best questions of the session came from Nick in the live chat: could LMS data be used to evidence the impact of accessibility, spotting where learners hit barriers and using that to drive improvements?

It’s close to Lee’s heart. He works with Blind Veterans UK, a charity supporting veterans who’ve experienced sight loss.

His answer: the data won’t hand you the answer directly, but it can point to the barriers.

It links straight back to PACE. If learning isn’t accessible, low participation or ability scores might reflect access problems, not motivation or capability.

There are a few patterns worth looking for.

Drop-off points, where lots of learners stop at the same page or activity. Lee’s nephew, who’s open about having ADHD and autism, switches off at a certain point in e-learning, and nothing after that lands. But it’s not just neurodivergent learners. “Years ago, it used to be we have a twenty-minute attention span. I’d argue that we’ve got a twenty-second attention span.”

Repeated failed quiz attempts. Sometimes it’s as simple as a drag-and-drop activity that someone’s accessibility software can’t cope with.

Time outliers. Someone taking much longer than average could be using a screen reader, working in a second language or dealing with a slow internet connection.

Lower completion among frontline workers, shift workers, certain regions or mobile users. 

That often points to bandwidth or time, not just accessibility. Devices and browsers can throw up surprises too.

Then look beyond the LMS. Support tickets and feedback often show the same complaints: the video won’t play, the course won’t navigate.

Lee’s rule is to start with the course, not the person. Use the patterns to find content to fix, test it against WCAG standards and ask learners directly. Then fix it and measure again.

Tom would take it further. Pick your most problematic content, the one with the dropouts and terrible quiz scores, and make it your target.

Fix it properly, then follow it through to the wider business impact. Ask people about the cultural difference it makes, too.

Retrospectively fixing content is hard to get buy-in for. Do it well once, and you’ve built your proof case for working accessibly from the start.

Where to start with proving L&D impact

If you’re only reporting completions and happy sheets right now, Tom’s first message is: don’t feel bad. “You are in the massive majority.” 

You’re not behind, and you’ve got solid data to start from.

Lee’s advice is to pick one programme. “It’s simple as that.”

Agree the effect with the business before you build absolutely anything.

Then make sure the data underneath is clean. 

“If the data is garbage in your system, it’s gonna mean absolutely nothing.” 

Around half the organisations Lee works with tell him their HR data isn’t 100% correct, so this step matters more than it sounds.

It also helps when learning and performance live in the same place. 

In Totara with BuildEmpire, people data from your HR system is the foundation. Participation comes through programmes, courses and certifications. 

Ability shows up in skills and competency frameworks. Change is evidenced through continuous performance management.

The effect needs the business’s own numbers, like sales or incident data. Report builder and APIs let you join those to your learning and performance data, without stitching spreadsheets together.

As Lee was clear to point out, no platform proves impact in isolation. It gives you joined-up evidence for the conversations that do.

And those conversations now reach further than ever. Lee’s speaking to IT, marketing and sales, not just L&D and HR. 

Ask your IT team what their BI tool can pull, then ask if it can sit next to your learning data.

Prove impact convincingly once, and the organisation will want it again. Then it just becomes routine.

Pull it all together, and the pattern repeats: start with the effect, not the activity.

Plan backwards from what the business wants, and report forwards from participation.

Bring managers in, because they shape whether learning turns into performance.

Collect more data than you think you need, and question anything that looks too good.

And turn it into one headline, in business language, with a clear next step.

Get that right, and L&D stops proving it was busy and starts proving it made a difference.

More from this series

This was the third of five sessions in BuildEmpire’s webinar series in partnership with Totara. Catch up on the first, Distributed Learning: Design for Connectivity, Not Proximity, and the second, Blended Learning Strategy: How to Design for the Learner.

Follow our LinkedIn page to register for the next session in January, where Tom will be joined by Mel Milloway to dig into the L&D tech stack.

And sign up for our newsletter to get more insights on learning and development just like this.

FAQs

What is L&D impact?

L&D impact is the measurable difference learning makes to how people perform and to the organisation’s results. It goes beyond activity measures like completions, attendance and satisfaction scores. Lee Cooper’s PACE framework breaks it into four stages: participation (what people did), ability (whether they can do it), change (whether they’re doing something different) and effect (whether the business improved). Proving impact means getting all the way to effect.

How do you measure the impact of training?

Start by agreeing the effect the business wants before you build anything, then plan backwards and report forwards. Set a baseline, compare trained and not-yet-trained groups, and track trends over time. Use the business’s own numbers rather than inventing new metrics, involve managers to evidence behaviour change, and look at the same story from several sources, like manager ratings, business metrics and learner accounts.

How should L&D report impact to senior leaders?

Lead with one headline that answers “so what?”, such as teams whose managers completed a programme had lower early attrition. Tell it in the business’s language, starting with the effect and using change, ability and participation as supporting evidence. Keep it to one page rather than a dashboard of charts, and finish with a clear recommendation for what to do next.