Why Satisfaction Scores Don't Prove Impact
High satisfaction and zero measurable impact can coexist — satisfaction is worth tracking, but it answers a much narrower question than 'did this work.'
What satisfaction scores actually measure
A satisfaction score measures how participants felt about an intervention immediately afterwards — engagement, relevance, quality of delivery. It's a genuinely useful signal about delivery quality. It is not a measure of whether behaviour, performance or risk subsequently changed.
Why high satisfaction and zero impact can coexist
People can enjoy an intervention, rate it highly, and return to exactly the same behaviour afterwards — enjoyment doesn't require change to have occurred. Satisfaction and impact are correlated in general, but weakly enough that one cannot substitute for the other in any specific case.
The seduction of an easy positive number
Satisfaction scores are fast to collect and reliably positive, which makes them an attractive headline metric — especially compared with behaviour or results data, which take longer to gather and are less guaranteed to be favourable. That asymmetry is exactly why satisfaction gets over-reported relative to its actual evidential value.
What to measure alongside satisfaction, not instead of it
Satisfaction remains worth tracking as a delivery-quality signal, alongside — never instead of — a measure of the actual outcome the intervention targeted. Reported together, satisfaction explains part of the story; reported alone, it's frequently mistaken for the whole of it.
Common questions on this topic.
No — they're a legitimate and useful measure of delivery quality. The issue is reporting them as if they answer the impact question, not their existence.
Yes — content people found difficult or uncomfortable can still be exactly what changed behaviour, which is another reason satisfaction alone is a poor proxy for impact.
Whatever specific outcome the intervention was funded to move — compliance gap closed, error rate reduced, time-to-competence — measured against a baseline, not a feelings-based proxy for it.
Want this thinking applied to your organisation?
Insight is useful. Applied insight changes outcomes. Let's talk about yours.