Do-Not-Track and the Economics of Third-Party Advertising | Hacker News Reader

Do-Not-Track and the Economics of Third-Party Advertising

papers.ssrn.com

27 pointscpeterso11 comments

The key assumption, I think, is here:

> Among the top 10,000 sites that show third-party capable advertising, we find that typically 15% of users visit the site at least 10 times per month. If one-fourth of such loyal users ultimately subscribe to the site, we estimate that a monthly fee of $2 would generate revenue comparable to that earned from third-party capable advertising, based on current ad rates.

Seems like quite an assumption that 1/4 of loyal readers will pay for a website. They cite the NYTimes paywall as evidence, but the NYTimes is hardly a typical site or a typical brand.

That's probably due to the 15 odd years of a lack of micropayments on the web, though. I'd pay $2 monthly for every one of the (very few, maybe 15-20) sites that I visit in return for privacy guarantees and no irrelevant advertising.

Unfortunately it's difficult to pull off a model where users can pay to turn off ads because it drastically reduces the per-user value of the remaining ad impressions.

Few people use ad blockers. Less than 7% of Firefox users have Adblock Plus installed:

https://addons.mozilla.org/en-US/firefox/compatibility/33.0

Funnily enough, I'm developing the technology for exactly this.

One of the coauthors here. 1/4 was actually a conservative number compared to what was found by the NYTimes paywall study. Unfortunately, we do not have more comprehensive studies to cite for a more generalizable number. Note however that the price one needs to pay according to our analysis is much lower than NYTimes (~ 2 dollars per month) and an average user would need only 2-3 such subscriptions.

The real question I have is: Do ads served based on a person's history generate more clicks/sales than ads based on the relevance to associated content, or even randomly selected? My subjective experience with targeted ads is poor - they seems to be ads for stuff I already bought and therefore, don't need anymore.

The only website that has ever shown me novel, relevant ads was Feedly. With all my RSS feeds, they have a centralized view of my interests.

Hi, co-author of the paper here. To answer your question, yes, there is evidence to suggest that targeted ads work better: http://papers.ssrn.com/sol3/papers.cfm?abstract_id=1600259

IMHO that's because ad targeting and ad profiles are still extremely crude. It sounds like you're talking about retargeted ads -- you look at a product and it seems to follow you around afterwards. The tech is too dumb to know you already bought it, but on average it's pretty effective.

Reply on news.ycombinator.com