Every L&D team fights the same two-front war. Stakeholders want the modules built quickly and don't want to discuss the underlying problem behind the request. Leadership wants proof the investment worked, using a budget that gets questioned first the moment things get tight.
Somewhere in between, the actual learner shows up to whatever gets built and decides in the first few seconds whether it's worth their attention.
Pocket Casts
Above is the conversation between Jess Almlie and Brad Batesole on why L&D keeps optimizing for the wrong things. Below are my takeaways on what to fix first, from the stakeholder conversation on day one to where AI actually belongs in your design process.
TL;DR
- Completion rates prove attendance, not learning, and stakeholders need to hear that difference up front.
- Learners abandon L&D content for the same reasons they abandon a paywalled article: too many steps and too much friction.
- Writing down expected outcomes before a project starts protects you and sets the real success metric.
- Budget requests land when they attach to a named priority leadership, not as a general case for more resources.
- AI can spot patterns in learner data, but the design decisions that fit your actual audience still need a human.
1. Your completion rate can hit 100% while nothing actually gets learned
Picture the mandatory compliance module everyone finishes by Friday. Nobody fails it. Nobody remembers it by Monday.
Brad sees the same pattern across every L&D team he's worked with. Optimize for cost first, completion second, volume third. Efficacy doesn't make the list, not because nobody cares about it, but because it's the hardest metric to defend the moment a budget gets cut. Cost and completion produce a clean number you can put in a slide. Behavior change takes months to show up and doesn't fit into a single report.
Which is exactly why completion survives when everything else gets cut. Think about the last time you clicked through a required course while a meeting ran in another tab, hunting for the "next" button before each screen finished loading. You finished. You'd swear under oath you learned nothing, and you'd be right. Brad makes the same point about the compliance work everyone powers through on autopilot. The box gets checked. The number goes up. Nothing about how that person does their job changes.
That's the tell. If nobody would voluntarily repeat a program, the program was built for completion, not learning, and the dashboard is only telling you that people showed up.
The fix isn't a new metric. It's naming the one you already have for what it is. A completion number is an attendance record. Say that out loud to the stakeholder asking for it, before the project starts, not after the results disappoint someone.
How to separate attendance from actual impact:
- Report completion as attendance in the same breath you report it, so a stakeholder never reads a 90% completion rate as 90% of the outcome they wanted.
- Set the behavior-change target before the content exists, because a specific "they should be able to do X on the job" written down on day one is the only thing a completion number can later be measured against.
- Add a lagging measure to every launch, a 60- or 90-day check on whether the behavior actually moved, since the gap between that number and your completion rate is the real story worth reporting.
- Treat a completion-only request as a decision, not a default, and make the stakeholder actively choose it once they've heard what it will and won't produce.
2. Learners hit one hoop and leave, exactly like they do everywhere else
You've clicked a link a friend sent, hit a subscribe wall, and closed the tab without a second thought. Employees do the same thing to your LMS.
Brad watched this happen constantly at Lynda.com. People abandoned sign-ups because the form asked seventy-two questions before anyone reached a single video. A survey would pop up asking something irrelevant before the content even loaded. The platform wouldn't work right on a phone, so people gave up before they started.
Which means the friction problem was never a training problem. It's the same UX problem consumer apps solved a decade ago, and L&D never got the memo. A product team treats a clunky sign-up flow as a revenue emergency and rebuilds it within a sprint. L&D treats the same friction as an IT ticket that sits in a queue for a year, because nobody's measuring what happens before someone reaches the content, only what happens if they finish it.
🔒 The remaining content is available exclusively to Pro Members. Click to unlock this full resource & much more
Picture someone opening your LMS on their phone during a ten-minute break. The login screen doesn't fit the display. They tap around twice, give up, and go back to whatever they were doing. Your completion report calls that "did not engage" and never asks why. All the investment you put into good instructional design is wasted if nobody gets past that screen.
Ways to strip friction out of the path to content:
- Time your own enrollment path on a phone, from the first click to the first second of real content, and treat anything over about a minute as a leak you need to fix before you build anything new.
- Delete every form field and survey that doesn't route the learner somewhere, since data you collect at the door costs you the people who close the tab rather than fill it in.
- Measure drop-off by step, not just final completion, because the screen where people quit tells you what to fix, and a completion rate only tells you that something is wrong somewhere.
