What kind of evidence would you want? The thing with social marketing is you can’t measure ROI very easily – with regular advertising you can usually link a new campaign with an increase (or decrease) in sales. With behavioural change, you don’t always have a clear baseline to measure from.
How many people accurately self-report their smoking habits in the first place? And how many will accurately self-report the progress of their quitting? And how can you determine whether they quit because of an advert or because of their GP or because a relative is dying of lung cancer or because their budget is too tight to afford cigarettes any more? It’s really complex. It’s easier with e.g. Army recruitment campaigns – did we recruit more good people than the year before? Then our advertising is working.
Obviously advertising needs to show some kind of ROI, otherwise there would be no point in doing it and departments wouldn’t be able to justify spending money on it. There is evaluation in place for social marketing, and it’s evolving all the time, but actually marketing campaigns aren’t always the catalyst for behavioural change. Usually you’re much more likely to change if, e.g. your peers don’t drink a lot than if you see a billboard telling you drinking is bad (substitute almost any other example here). An advertising campaign is something that sits alongside that to support you and normalise the change, but it’s pretty rare anyone successfully alters their behaviour based on an advert.