You are reading a single comment by @ewanmac and its replies.
Click here to read the full conversation.
-
Hadn’t seen that Shapiro paper but a colleague of mine is working on what seems like a very similar paper. Massive dataset, sophisticated analysis. Conclusion is that for some types of advertising the uplift is v. small but safe whereas for other types it’s high risk high reward. He presented it to a group of CMOs recently and they said this fit their anecdotal observations.
I’ll be sure to post the paper here in the memes thread when it escapes review.
ewanmac
There's a really interesting recent paper which looks at TV advertising for some of the biggest consumer-packaged-goods brands and the findings challenge not just the conventional wisdom that 'advertising works' (otherwise why would companies spend billions on it!?) but also a backlog of previous marketing literature suggesting it was fairly effective at boosting sales. Previous literature estimated ad elasticities of 0.15-0.2 (so if you double your ad spending you increase your sales by 15-20%), while this new paper finds an ad elasticity of 0.01. New study is substantially superior in methodology and sample size. Combining the ad spend of these brands with these elasticities gives negative ROIs for over 80% of the brands in their sample, implying that for those brands reducing advertising would increase their profits!
The authors wondered why it was that firms spent all this money on advertising if these results were 'correct'. They had two interesting views. (1) while these firms are aggressively profit maximizing entities, the advertising manager at the firm may have skewed incentives; if they did lots of digging into effectiveness they might put themselves out of a job. (2) it's very hard to measure effectiveness and simple empirical methods that are readily understood by corporate types won't account for the massive endogeneity problem that you are dealing with ie firms advertise more during periods of high demand - if you don't account for the fact that demand would naturally be higher during certain periods when firms tend to advertise more, and you falsely attribute that increase in sales to the causal effect of advertising, that could lead you to overstate the effect of ads. So you need a clever approach as part of an econometric analysis to deal with this.
Big qn I guess is this research seems to suggest a lot of TV advertising is wasteful but digital - ie targeted - advertising might be much more effective...
Paper: https://papers.ssrn.com/sol3/papers.cfm?abstract_id=3273476