Four levels, sized for reality
The classic ladder, reaction, learning, behavior, results, survives translation to SMB scale. Reaction: one question, would you recommend this to a colleague in your role, not a satisfaction survey. Learning: a real check, a simulation passed, a work sample reviewed, not a quiz of definitions. Behavior: the observable change, are the one-on-ones happening, are the proposals following the structure, checked at thirty and ninety days. Results: the operational metric the program targeted, moving in the right direction.
Each level costs more to collect and proves more; the discipline is climbing as high as your evidence honestly reaches and no higher.
Behavior is the honest middle
Results attribution is genuinely hard, too many variables touch revenue, but behavior is observable by any manager: the skill either shows up in the work or it does not. Define the two or three observable behaviors per program before it runs, who will look for them and when, and collect the observation as part of normal management rhythm, not a research project.
Behavior evidence also localizes failure usefully: taught but not practiced indicts the program design; practiced but not helping indicts the skill choice; both fixable once distinguished.
Connect to the metric, report without pretense
Every program was aimed at something operational, error rates, ramp time, win rates, review scores, retention of the trained population, or it should not have run. Track that metric before and after, report the movement alongside the behavior evidence, and say plainly that other factors also moved: leadership trusts modest claims with evidence over precise claims without.
The quarterly learning report is one page: what ran, who completed, what behaviors changed, what metrics moved, what the next quarter changes as a result. That last line, the program learning from its own evidence, is what makes the budget renew.
Progress, behavior and the report on one screen
The Learning Desk tracks completion and assignment evidence per learner, keeps the readiness criteria as the competence check, and Progress and Reports assembles the quarterly picture with a Velora leadership report on top, adoption, completions, risks and next steps, in plain prose.
Because the desk shares the workspace with the operational metrics the programs target, before-and-after is a look, not an export. From the Plus membership.
Frequently asked questions
How do you measure training ROI in a small company?
Climb the evidence ladder honestly: real competence checks, observable behavior change at thirty and ninety days, and movement in the operational metric the program targeted, reported with modesty.
What is the best indicator training worked?
Behavior: the skill visibly appearing in the work, defined before the program ran and checked by managers in normal rhythm. It is collectable and hard to fake.
What should a learning report to leadership contain?
One page: what ran, completion, behaviors changed, metrics moved, and what changes next quarter as a result. Modest claims with evidence renew budgets.