3 ms·
I don't know about 'authoritative', but ads do seem to hurt. The existing experimental and quasi-experimental literature shows very bad effects on user behavior
by gwern 8y ago
I don't know about 'authoritative', but ads do seem to hurt. The existing experimental and quasi-experimental literature shows very bad effects on user behavior:
1. "Measuring Consumer Sensitivity to Audio Advertising: A Field Experiment on Pandora Internet Radio", Huang et al 2018: https://davidreiley.com/papers/PandoraListenerDemandCurve.pdf https://davidreiley.com/papers/PandoraListenerDemandCurve.pd... [experimental]
2."The Effect of Ad Blocking on [Firefox] User Engagement with the Web", Miroglio et al 2018: https://research.mozilla.org/files/2018/04/The-Effect-of-Ad-Blocking-on-User-Engagement-with-the-Web.pdf https://research.mozilla.org/files/2018/04/The-Effect-of-Ad-... [quasi]
3. and my own A/B test: https://www.gwern.net/Ads https://www.gwern.net/Ads [experimental]
- ta3216 8y agoI think it depends on the audience. Looking at your blog I can see how ads may not work as well. My blog is focused more on dining out, shopping etc where viewers think nothing of ads and I think they would even think it weird if there weren’t ads.
- gwern 8y agoYou could say that about both Pandora (radio is expected to have ads) and Mozilla's general Internet sampling frame (the Internet is expected to have ads), but they still find a large average harm.
- ta3216 8y agoMy site is focused on shopping and dining out therefore I don’t think the conclusions in those papers apply in my case. The audio interruptions noted in the Pandora test are far more intrusive than in-page display ads. If I used interstitials then I would agree.