A new benchmark on how organisations run learning has just landed, and one figure stands out for anyone who has to defend a training budget. According to TalentLMS, in its 2026 L&D Benchmark Report, only 37% of companies measure learning and development by business impact. The single most common way to judge a programme is still whether people finished it and whether they liked it. That is a measurement gap, and it is hiding the only question that matters: did behaviour actually change.
TalentLMS found in its 2026 L&D Benchmark Report that only 37% of companies measure training by business impact, while 28% still measure success by how satisfied learners were. Completion and satisfaction are reaction-level metrics: they tell you a course ran, not that behaviour shifted on the job. Measuring impact means tracking observed behaviour after the programme, not feedback on the day.
Why this matters now
Budgets are under sharper scrutiny than they have been for years. The same TalentLMS report notes that the share of executives who see L&D as a cost rather than an investment has fallen from 54% in 2022 to 41% in 2025, which is real progress. But that goodwill comes with a condition: leaders now expect evidence. If the only proof a programme worked is a completion percentage and a satisfaction survey, the case collapses the moment a finance director asks what changed because of it. That gap is exactly where training budgets get cut first.
What the research says
The numbers in the 2026 report describe an industry measuring the wrong end of the process. Business impact is selected by 37% of HR managers as a measure of L&D success. Satisfaction with training is selected by 28%. In other words, more than a quarter of organisations still treat how learners felt about a session as a headline indicator of whether it worked. Feeling and impact are not the same thing, and the distinction has been understood for decades.
Kirkpatrick and Kirkpatrick (2016) set out four levels of training evaluation: reaction, learning, behaviour and results. Reaction is whether people enjoyed it. Learning is whether they can recall the content. Behaviour, Level 3, is whether they actually do anything differently at work. Results is the business outcome that follows. Satisfaction scores and completion rates sit at Level 1. The 37% figure tells you how few organisations are climbing past the bottom rung to the level where genuine behaviour change can be seen.
There is a deeper reason self-report is weak evidence. Sidestream's own academic behaviour-change work, building on research from UCL, Cambridge and Bocconi, found that participants rated their own skill highly while their measured performance told a more honest story. People are confident judges of how a session felt and poor judges of how much they changed. A measurement system built on asking them is built on the least reliable instrument in the room.
What most organisations do (and why it fails)
The default sequence is familiar: book the training, run the training, send the feedback form, file the completion data. The numbers look healthy because completion and satisfaction are easy to collect and almost always come back positive. People are polite, and a day out of the inbox tends to score well. None of it tells you whether a single manager handled a difficult conversation differently the following Monday.
This fails for a structural reason, not a lazy one. Level 1 and Level 2 are cheap and immediate; Level 3 is harder, slower and requires you to define the behaviour you wanted before the programme starts. So organisations measure what is convenient and quietly assume the rest. The result is a portfolio of training that everybody enjoyed and nobody can prove worked.
What works
Some things cannot be taught, they have to be felt, and the same is true of proof. You measure behaviour change by deciding, before anyone is trained, what observable difference you expect to see, then looking for it in the weeks after. That means defining the specific behaviours, capturing a baseline, and using behavioural observation rather than a satisfaction form to check whether the shift held. It is more work than counting completions, and it is the only approach that survives contact with a sceptical CFO. We set the method out step by step in our guide to how to measure behaviour change, and we go deeper on the financial case in our piece on measuring the ROI of behaviour change training.
This is also why the design of the training matters as much as the measurement. A programme built around lived practice gives you something observable to measure in the first place, because participants have already performed the behaviour under realistic conditions rather than just discussed it. That is the foundation of behaviour change training done properly: build for the behaviour, then measure the behaviour, instead of running a session and hoping the survey speaks for it.
Completion tells you a course ran. Satisfaction tells you it was pleasant. Neither tells you a single person did anything differently the following week.
Where Sidestream fits
We are a behaviour change consultancy that combines organisational psychology with immersive theatre, and measurement is built into how we design, not bolted on at the end. Our work spans workshop-scale training, larger immersive simulations and tailored behaviour change programmes, and every one of them is built around a behaviour you can name and observe. If your last programme produced great feedback and no proof, book a free 30-minute diagnostic call and we will talk through what to measure instead.
The 37% Gap: The Takeaways
TalentLMS found in its 2026 L&D Benchmark Report that only 37% of companies measure training by business impact, while 28% still rely on how satisfied learners were. Completion and satisfaction are reaction-level metrics that tell you a course ran, not that behaviour shifted. Measuring impact means defining the behaviour in advance and observing it afterwards, which is the only evidence that survives a budget review.
- TalentLMS (2026 L&D Benchmark Report) found only 37% of companies measure L&D by business impact; 28% measure by training satisfaction.
- The same report found executives seeing L&D as a cost rather than an investment fell from 54% in 2022 to 41% in 2025, so the demand for evidence is rising.
- Kirkpatrick and Kirkpatrick (2016) place behaviour at Level 3; satisfaction and completion sit at Level 1, the rung most organisations stop at.
- Active retrieval raised retention by around 50% over re-reading (Roediger and Karpicke, 2006), so programmes built on practice give you something durable to measure.