13 Mar 2024 Search

User research about advertising, personalization, and privacy is surprisingly consistent. In a replication prediction market, I would invest in futures on any research that shows:

  • About 30 percent of people prefer personalized ads to non-personalized ads.

  • About 30 percent of people prefer non-personalized ads to personalized ads.

  • For the other 40 percent it depends how you ask.

A lot of good work has been done in this area, but the results are inconvenient for anybody who wants to be able to build one set of online advertising software and settings for everybody. (This is separate from the question of whether personalized advertising is good for you or not—it’s not. People have the right not to be protected from ad personalization, if that’s really what they want.) Research links, in order of percentage of pro-personalization people found.

18%

McDonald et al. did in-person interviews with questions on people’s knowledge and preferences about Internet ads.

  • 18% agree with Glad to have relevant advertisements about things I am interested in instead of random advertisements

  • 30% agree with No one should use data from Internet history to personalize ads.

26%

Accountable Tech did a 1,000-person survey.

  • 26% agreed with I’d rather see relevant ads, even if companies are using my personal data to target them.

  • 46% supported a ban on collecting people’s personal data and using it to target them with ads.

30%

Verbraucherzentrale Bundesverband commissioned a phone survey of 1,500 people in 2024. Seven out of ten respondents (70 per cent) agree that companies should not be allowed to create profiles based on personal data to deliver personalised advertising.

30% (or 47%)

U.S. Internet Users Ready to Limit Online Tracking for Ads In a Gallup poll of 1,019 adults, 30% agree that advertisers should be allowed to match ads to interests based on websites visited but 47% want to be able to allow advertisers they choose to personalize ads to them in some way. (It depends how you ask.)

32%

Turow et al. did an online study of 314 people.

  • 32% said yes to Please tell me whether or not you want websites you visit to show you ads that are tailored to your interests.

  • 68% definitely would not allow personalization based on web history even if anonymous.

36%

YouGov did a survey of British adults.

  • 36% are Personalized Pioneers who say they want personalized ads
  • 55% say that personalized ads creep them out.

37% (26% + 11%)

In a study of regular social media users, YouGov found that 11% are fine with personal data for all ads, and an additional 26% only object to personal data being used for political ads.

42%

IAB Europe did a survey of 11,020 European Internet users.

  • 42% don’t mind personalized ads based on browsing data

  • 20% don’t mind having data shared with third parties for ads

People are different

Some people are privacy people and I’m one. Personally I don’t understand why anybody would want a personalized ad at all. I’m most likely to get any use from advertising when I need to buy unfamiliar stuff, and that happens when I’m learning about a new activity that I’m not good at. I want to see the same ads that show up for people who already know the skills and the scene, and would raise a stink about a deceptive ad in their information space. As far as I can tell, personalization makes advertising work in the wrong direction.

Some people are personalization people and don’t understand why you wouldn’t want a personalized ad. If you’re going to look at an ad anyway it might as well be for something that you’re more likely to buy. (But I want to see the same ad that a company is willing to show to the regulators, editors, and experts in their community of practice, not what they think they can get some random person to click on. I don’t understand Kevin and he doesn’t understand me.)

Other people, the biggest group at 40 percent, are probably better off not overthinking online ads and, instead, learning about other stuff.

Whatever a person prefers,Can database marketing sell itself to the people in the database? it seems to be about the personalization using any information from outside the context in which the ad appears—cross-context behavioral advertising—not about individualized tracking. Jerath et al. found similar perceived privacy violations (PPV) for new ad personalization technologies that prevent individual identifiers from being used: New technologies or proposals that ensure that data are kept on the consumer’s machine lower PPV relative to behavioral targeting but, importantly, this decrease is small. Furthermore, group-level targeting does not differ significantly from individual-level targeting in reducing PPV. The IAB and Gallup studies imply that there is a cohort of users who want personalization but not cross-context tracking, but safer personalization looks like a small add-on to the 30%, not a widely held preference that splits off a chunk of users from the core anti-personalization group.

using the 30-40-30 rule

A lot of this research looks useful as a way to spot deceptive patterns in personalization preferences. If fewer than 30% or more than 70% of users end up with ad personalization turned on, something is probably wrong with the personalization UX or the underlying trustworthiness of the medium.

And when building anything that depends on advertising, it’s going to be hard to build something that only works if everybody who uses it has personalization, or that only works if nobody does. As far as I can tell, adding extra personalization for those who don’t want it just builds support for privacy habits, settings, tools, and laws, in that order. And banning personalization entirely, or building a medium that can’t be tweaked to support it, hasn’t really been tried yet but would probably cause problems of its own.

In my humble opinion, the biggest benefit of the 30-40-30 rule is that it helps justify the decision to support Global Privacy Control everywhere. People are just different, so let everybody pick once whether they want cross-context behavioral advertising everywhere. Yes, I know, I know, under a different set of laws we could have had Global Personalization Control, that would work the other way around so that the work of turning it on would be done by the personalization people and not the privacy people. But GPC is there, it works, and if sites, browsers, and platforms can implement it in a common-sense way you should just be able to turn it on once and be happy for as long as you own the software or device, kind of like Filmmaker Mode on a TV. In the long run, I can safely predict that the fraction of users who turn GPC on will be somewhere between 30 and 70 percent. Which end of the range we end up with will depend in how trustworthy the ads, the Internet, and the economy are in other ways, but that’s another story. More: GPC all the things!