Scaling failures are not failures of ambition. The funding and the political may be there, the pilot worked. What goes wrong is further down the line: the parts that get edited out, usually the parts that made the intervention work in the first place. This happens through a series of reasonable-looking compromises, and by then, the intervention has been ‘scaled’ but its key ingredients have not.

Entertainment changes behaviour through mechanisms we understand well. Narrative persuasion lowers cognitive defences, so ideas enter without the resistance a health message would trigger. Identification maps audiences onto characters. Modelling means we learn vicariously from watching consequences play out on someone else. Emotional arousal cements the storyline in memory.

When researchers tracked what happened after girls in Uganda watched Queen of Katwe, the 2016 film based on Phiona Mutesi, who picked up chess at nine in a Kampala slum and went on to compete internationally, the girls who saw it went on to perform measurably better in their exams (Riley, 2022). One film. A measurable shift in educational attainment. You can see why the sector loves this.

I recently spoke with practitioners working at the leading edge of social and behaviour change programme design. Five organisations spanning India, East Africa, West Africa, Latin America and the UK, each with very different operating models and funding structures. Five unrelated conversations surfaced the same tensions, and those tensions tell an interesting story about where this field is. Here is what I took away.

Every few weeks, a new AI tool lands in our inboxes with the implicit message: adopt this, or fall behind. For social impact leaders, the pressure is intense. Funders are asking about it and teams are experimenting with it without a clear overview of what value it is adding (if any). The question is not “which AI tool should we use?” but “what does good AI adoption actually look like for us?”. That is a behavioural question.

If we’ve ever chatted about research and reporting, you know my gripe about evaluation reports that are never revisited. It is a missed opportunity to learn something genuinely useful about whether your organisation is actually changing anything.

Across the impact sector, evaluation is often treated as a reporting obligation rather than a strategic tool. The result is a sector that spends significant resources measuring its work without fundamentally improving it. That cycle can be broken. But it requires starting with a different framing.