# How Do We Know It Worked? Measuring L&D Impact How Do We Know It Worked? A Practical Guide to Measuring the Impact of Learning | BridgeTheir Shared Guidance Start with a business outcome and use evidence the business already trusts. Introduction ### We May Be Starting With the Wrong Question L&D often walks into a budget conversation already on the back foot, ready to defend its work before anyone has asked a question. That may be the first thing worth reconsidering. Ask why it is so difficult to justify an investment in developing people, and you'll hear some version of the same answer from many: L&D may be answering a question nobody asked. A CEO wants to see something move—a clearer line to productivity, revenue, retention, or another business priority. A description of a well-designed course does not necessarily make that connection visible. Even the usual framing can lead us in the wrong direction. "How do I prove ROI?" assumes the answer, then looks for evidence to support it. That can lead to some strange mistakes. "How do I find out whether this is working?" on the other hand, leaves room for an honest answer. The willingness to discover that something didn't work is also what makes a positive result more credible. There's also a more forgiving way to read why the business asks for proof at all: it isn't an accusation. Every function that spends the company's money gets asked to show its value in numbers—sales, marketing, operations. L&D isn't being singled out as suspect. It's being treated like every other cost center, which, uncomfortable as it feels, is actually fair. We spoke with four people who've spent years inside this exact argument: an instructional designer who has built programs for thousands of employees, a talent development consultant who has worked across the globe, an author and broadcaster who has interviewed hundreds of L&D leaders, and a consultant who has spent his career pushing L&D to adopt evidence-informed practice, and none of them recommend a better ROI formula. All four, independently, point to the same shift: the case becomes stronger when it starts with a business outcome rather than a course and uses evidence the business already trusts. This playbook brings their perspectives together, along with practical ways to connect learning to outcomes the business already recognizes. Part One ### Why the Usual Metrics Aren't Enough Completion and login data is useful operationally—it just isn't the whole case.Ask what is least useful in a leadership conversation, and completion rates and satisfaction scores come up almost immediately. A 98% completion rate or a four-star course rating may matter operationally, but neither tells a budget owner what actually changed. We chatted with Moe Ash, founder of an instructional design consultancy who explained that part of the problem is the label itself. There's a useful distinction between return on *investment* and return on *impact*. The first asks whether money was well spent in the abstract, the second asks whether something specific actually changed as a result. A completion rate answers neither question. It just confirms a box got checked. > "It doesn't really represent learning. It represents a checkbox." L&D teams often stumble when they treat training as the fix for what's actually a systems or process problem. If there's no clear channel for a career conversation, no time for a manager to have one, or no shared understanding of what growth even looks like at a given company, a course won't touch any of that—and measuring its completion rate may look like success while the real problem stays exactly where it was. This doesn't make completion or satisfaction data worthless. It simply means those measures cannot carry the whole story. The more useful question is not only "did people finish it?" but "did anything change that the business would recognize as meaningful?" #### Signs the Measurement May Not Be Telling the Whole Story - The metric you'd report (completion, satisfaction, seat time) has no obvious link to a number your CEO already tracks. - Training is being discussed before anyone has confirmed whether the underlying issue is a process, system, or access problem. - Nobody outside L&D would recognize your success metric as evidence of anything. - The conversation is focused on return on investment when impact—a specific behavior or outcome—may be the more useful measure. Part Two ### A Different Question Can Lead to Better Evidence Before L&D can measure whether people are engaged, or whether a program changed anything, it helps to understand what is actually happening in the organization. That means **shifting from "I want to *prove* my value" to "I want to find out whether this learning initiative *creates* value."** Start from the gap the business can see, not from the course catalog.We spoke to Elena Agaragimova, a talent development consultant who has worked with organizations across the globe. She sees the same pattern show up almost everywhere. The first signs of employee disengagement tend to show up outside the learning program itself. Are employees taking career questions directly to HR because they don't know where else to go? Are more people quietly updating their LinkedIn profiles? Has performance or promotion activity gone flat? And underneath all of it, do managers feel equipped to have development conversations, or is that the real gap? > "Everything streams from the managers. Do they know what career paths look like?" These signals could point to very different problems. Employees may not see opportunities to grow. Managers may need more support. A process may be getting in the way. Training may be part of the answer, but the same course will not solve all of them. This kind of diagnosis matters even for training that rarely gets questioned, including compliance. Compliance training is often the most visible program L&D delivers, reaching nearly everyone—including the executives deciding its budget. Look for signals outside of completion rates: Are people making fewer mistakes, raising concerns earlier, and working together more consistently afterward? Having this information gives a skeptical executive clearer evidence that the training is influencing behavior, strengthening collaboration, *and* protecting the business. #### Questions Worth Exploring - Are employees participating and engaging with the training, and what happens after they do? - Are people taking career questions directly to HR because they don't know where else to go? - Is there a visible increase in employees quietly job-searching? - Has performance or promotion activity stalled? - Do managers feel prepared to have development conversations, or is that the real gap? - For routine programs such as compliance, who is most at risk, and what matters beyond completing the course? Part Three ### What Fieldwork Can Tell Us One idea came through particularly clearly in our conversations: L&D could benefit from more fieldwork. We spoke to Nigel Paine, an author and broadcaster who has interviewed hundreds of L&D leaders on Learning Now TV, and he puts it bluntly: > "L&D does not do enough fieldwork. When someone says they need [something], you investigate. Can we talk to people? Can we go out there and understand, profoundly, where the blockages are?" Feedback from the people closest to the work is evidence, not anecdote.In the L&D context, fieldwork means treating an assumption as something to check before building around it. When someone reports a problem, there is value in speaking with the people involved, spending time with the team experiencing it, and finding where the blockage actually sits before proposing a response. What comes back may point to an action plan rather than a course. Informal support and confidence can matter as much as formal learning. Some issues described as skill deficits are really confidence deficits: people may know what to do but not yet believe they can do it. There's a named method worth borrowing directly here: the Brinkerhoff Success Case Method. Instead of interviewing everyone, the method advises interviewing the extreme cases—the clearest failures and the clearest successes—and compares the conditions around each, not the content. The finding, almost every time, is that failure has less to do with the course itself and everything to do with context: a manager who quietly blocks progress, a toxic environment, a process nobody's allowed to question. Fix the context, and the same content that failed on one team can succeed on another. Who decides what counts as evidence matters as much as the interviews themselves. L&D can ask leadership directly, "If you saw this specific change, would that convince you?" The measurement plan can then work backward from that answer. It is a small change in sequence, but it makes the final evidence much more credible because the organization helped define it. #### Tips on Structuring a Fieldwork Study - Starting with one team or function with a known, visible problem can offer a more useful starting point than a broad survey. - Separating interviews with the clearest success and failure cases can reveal differences in conditions, not only content. - Leadership can help define what evidence would convince them. - Anonymizing direct quotes can protect sensitive details and encourage honesty. Part Four ### Managers Can Help Make the Change Visible Fieldwork produces evidence L&D can bring back to leadership. There's another source of evidence that's easy to overlook because it doesn't look like data at all: the manager who was actually there to see the change happen. We spoke to Tom McDowall, founder of an L&D consultancy, about pushing back on the instinct to design *around* managers rather than *through* them. > "Line managers play a huge role in whether or not someone develops a skill. Thinking that we can replace them or go around them is just provably false, it doesn't work." A recurring 1on1 turns the manager into a witness rather than a distant stakeholder.A few relatively low-cost choices can help a manager become a direct witness rather than a distant stakeholder. Their presence at a launch or orientation means both manager and participant understand what the program is intended to change. A midpoint conversation gives them a reason to discuss what is being learned. If the manager is also part of the closing moment, they have an opportunity to notice—or not notice—whether anything changed. That last point connects to a simple measurement technique: participants rate their own skill level retrospectively, looking back from the end of the program. A straightforward before-and-after survey can inflate the starting point because people do not yet know what they do not know. They may rate themselves too highly, which makes the eventual gain look smaller. A retrospective rating helps reduce that distortion. The same question can also go to a participant's manager and direct reports: not only "did you feel this changed?" but "did the people around you notice it too?" None of this works if managers are simply handed a new expectation on top of everything else they're doing. The more workable path is to find out how a manager already operates—their existing rhythms, their existing check-ins—and fit the request inside that, rather than asking them to adopt an entirely new process alongside their current one. Resistance from a busy manager is rarely really about time; it's about being handed one more disconnected thing to manage. #### How Managers Can Become Witnesses, Not Just Stakeholders - Manager participation at the beginning and end of a program can create context and make their observations at the close a useful data point. - At least one touchpoint can give the manager and participant a reason to discuss the learning, not only its completion. - The participant, manager, and direct reports can answer the same retrospective question so their perspectives can be compared. - Before introducing something new, consider how the manager already works and whether the new expectation can fit into an existing rhythm. Part Five ### Where AI Can Help With the Evidence AI surfaces the pattern. People still decide what it means.AI is often discussed in L&D as a content shortcut: faster slides, quiz questions, and first drafts. Those are real uses, but they may be less valuable than AI's ability to recognize patterns across messy, high-volume qualitative data. A collection of interview notes, for example, can be analyzed for recurring issues, common blockages, and quieter themes that may be hard to spot when transcripts are read one at a time. Human judgment still matters. AI can help surface the signal; people still have to decide what it means. It's easy to get excited about a new capability. But if an AI coaching tool is introduced before the team's needs are clearly understood, it becomes much harder to tell whether it actually made a difference. The organizations getting real value from AI are using it to process evidence they already went and gathered, not as a substitute for gathering it in the first place. #### Where AI Can Help Build the Case - Synthesizing themes across dozens of fieldwork interviews faster than a person could read them all. - Flagging a trend inside performance or engagement data that wouldn't be visible manually. - Supports the fieldwork conversation (never entirely replaces it). AI can process what people share, but it cannot recreate the context of the conversation itself. Part Six ### Three Examples of Learning Creating Visible Change Case One A defect rate traced back to the factory floor Case Two A contact center that fixed a conversation, not a skill Case Three A rollout employees started quoting back at work #### The Manufacturer That Traced a Rising Defect Rate to the Floor A medical device manufacturer that makes replacement heart valves watched the defect rate on one production line climb well above the factory average. Quality assurance was catching the problem before any valve shipped, but the cost was already adding up: the parts are expensive and a delayed shipment could mean a patient went without a device they needed. Tom McDowall was brought in for his training expertise. Rather than building a course, his team first spent time on the floor observing the line and talking to the people running it—crucial fieldwork that ended up revealing an environmental cause. > "We'd go out onto the floor, we'd take up people's time. We'd really assess what was happening." In examples like Tom's, once the problem is pinpointed the appropriate training can be delivered. Training ROI is evident to leadership as L&D did its homework first. #### The Contact Center That Fixed a Conversation, Not a Skill An agenda before the conversation and a summary after it—two habits, not a workshop.At a contact center, a fortnightly employee happiness survey showed a clear dip tied to one recurring moment: the Friday-afternoon one-to-one between managers and the people they supervised. Tom McDowall was brought in to look at why training wasn't translating into the work itself. He didn't start by building a communication skills course. He started by observing the conversations already happening and asking what the strongest managers on the floor were doing differently from everyone else. > "They were making sure an agenda went out before the call, so people knew exactly what they were walking into. And they were following up afterwards with a short email outlining what was said and what support was available." Those two habits—a template agenda and a follow-up email—turned out to be nearly everything the strongest performers were doing that the rest weren't. Rolling them out, instead of running a two-day workshop or something similar, moved the happiness score back up within weeks. Call times and customer service scores recovered, and revenue collection finished about 4% above its previous baseline. This proves why L&D must provide wrap-around support—instead of delivering training as a one-and-done—to ensure learning drives real ROI. #### The Story Employees Started Telling in Their Everyday Work A company needed to roll out a new awareness model to 3,000 employees, but unfortunately most of those employees didn't even know the old model existed, let alone that it was changing. Moe Ash was brought in to consult and took a creative approach. Rather than a mandatory course, the rollout became a three-month story, delivered as mobile push notifications timed to when people actually had a spare moment. > "We used a mobile-based learning management system so we can do push notifications on their phones. And we managed to make the push notification come up in two different instances in the week. One was on Monday and the other was on Thursday. Monday during lunch breaks because that was the time they will be open up to something like that. And Thursday during commute as they're going back home." A recurring cast of characters worked through the new model episode by episode. People started looking forward to the next installment the way they would a streaming series. A month after the campaign wrapped, the clearest proof of impact showed up somewhere nobody was measuring directly: leaders and supervisors started using the story's own language and characters to talk about the awareness model because it had actually become part of how the organization talked to itself. #### What These Cases Have in Common - The teams looked closely at the problem before deciding what kind of support to provide. - The support fit into people's actual work, whether that meant a one-page template, a short follow-up email, or a mobile episode during a commute. - The outcomes showed up in metrics the business already tracked—a defect rate, a happiness score—not only in participation data. - The evidence came from people close enough to see the change: line workers, contact center managers, and employees. Not solely from an L&D report. - The results gave L&D a more credible story to bring back to the business than completion rates alone could provide. Part Seven ### A Practical Toolkit Use any of these ideas that fit your current question, available data, and level of measurement maturity. #### Retrospective Self- (or Peer) Assessment At the end of a program, participants can rate their current skill level and then look back at where they started. This avoids some of the baseline inflation that can happen in a traditional before-and-after survey. Asking managers and direct reports the same question adds useful perspective. When all three groups see a gain, the finding becomes much more persuasive. #### Root-Cause Diagnostic Checklist Before design begins, it may help to look for career questions escalating directly to HR, a rise in people updating their job profiles, stalled performance or promotion activity, and whether managers feel equipped to have the conversation at all. The same type of diagnostic thinking can be useful for compliance training, not only skills programs. #### Fieldwork Interview Guide The Brinkerhoff Success Case Method focuses on the clearest success and failure cases. Interviewing the two groups separately can reveal differences in the conditions surrounding them, not only the learning content involved. Leadership can also help define what evidence would be convincing before the measures are chosen. #### "Would This Convince You?" Script Before a measurement effort begins, ask your executives one direct question to clarify the goal: "If you saw this specific change, would that convince you it is working?" The answer gives L&D and leadership a shared definition of success around which the measurement plan can be built. Closing Thought ### Go From Proving to Discovering The four people behind this playbook did not make a stronger case for L&D simply by getting better at ROI math. They asked questions they were willing to be wrong about, looked for answers rather than assuming them, involved managers and comparison groups early, and returned with evidence the business had already agreed would be meaningful. And remember that not every tested idea works. You may run an experiment that makes no difference at all. Money that would have gone toward scaling it went somewhere else instead. It's a short story, but it's the sharpest illustration of whether something creates value. About Bridge ### Bring Learning, Skills, and Performance Together Learning has more impact when it stays connected to what happens next. Employees need opportunities to use what they've learned, managers need a clear view of where support may be helpful, and L&D needs a way to understand whether the program made a difference. Bridge brings learning, skills, and performance together in one platform. That makes it easier to support many of the ideas in this guide—from giving managers a more active role in development to looking beyond completion rates and seeing how employees are growing over time. The conversations still belong to employees and managers. Bridge simply helps keep the learning, development goals, feedback, and performance information connected, so everyone has a clearer place to start.