04 May 2024 Search

This is a follow-up to a previous post on how a majority of US residents surveyed are now using an ad blocker, and how the survey found that privacy concerns are now the number one reason to block ads.

Almost as long as Internet privacy tools have been a thing, so have articles from personalized ad proponents telling us not to use them, because personalized ads are good actually. The policy debate over personalized (or surveillance, or cross-context behavioral, or tracking-based, or whatever you want to call it) advertising seems to keep repeating an endless argument that on the one hand, personalized advertising causes some risk or cost, I’m not going to summarize the risks or costs here, go read Bob Hoffman’s books or Microtargeting as Information Warfare for more info but on the other hand we have to somehow balance that against the benefits of personalized advertising.

Benefits? Let’s see them. Cross-context behavioral advertising is good for consumers should be straightforward to test. If ad personalization really helps match buyers and sellers in a market, then users of privacy tools and privacy settings must be buying worse products and services. Research should show that the more privacy options you pick, the less happy you are with your stuff. And the more personalized your ad experience is, the more satisfied of a customer you are. This is different from asking whether or not people prefer to have ad personalization turned on. That has been pretty extensively covered, and the answer is that some people do, and some people don’t. This question isn’t about whether people like personalized ads or not, it’s about whether people who get more personalized ads are happier with how they spend their money.

This should be a fairly low-cost project because in general, the companies that do the most personalized advertising are in the best position to do the research to support it. Are users of privacy tools and settings more or less satisfied with the products and services they buy than people who leave the personalized ad options on?

  • Do privacy-protected users give lower ratings to the products they buy?

  • Do privacy-protected users return or stop using more of their purchases?

  • Are privacy-protected users more likely to buy a replacement, competing product after an unsuccessful first purchase in a category?

  • Are privacy-protected users more likely to agree with general statements about a decline in quality and trustworthiness in business in general?

The correlation between more privacy and less satisfied consumer would be detectable from a variety of angles. Vendors of browsers with preferences that affect ad targeting should be able to show that people who turn on the privacy settings are somehow worse off than people who don’t. Anti-adblock companies do research on ad blocker users—so how are shopping experiences different for those users? Any product that connects to a server for updates or telemetry is providing data on how long the buyer chooses to keep using it. And—the biggest opportunity here—any company that has an Apple iOS app (and that’s a lot of companies) should be able to compare satisfaction metrics between customers with App Tracking Transparency (ATT) on or off.

Ad platforms, search engines, social network companies, and online retailers all have access to the needed info on ads, privacy settings, locations, and purchases. Best of all, they’re constantly running customer surveys and experiments of all kinds. It would be straightforward for any of these companies to run yet another user satisfaction survey, to prove what should be an obvious, measurable effect. I’m really looking for any kind of research here, whether it’s a credit card company running a SQL query on existing data to point out that customers with iOS app tracking turned off have more chargebacks, or a longer-term customer satisfaction study, anything.

looking at the data we do have

The Consumer Welfare Effects of Online Ads: Evidence from a 9-Year Experiment | NBER Facebook users who had ads turned off entirely were asked to place a monthly valuation on Facebook, and the amount was not significantly different from the valuation given by users with ads turned on. (The researchers did not test personalized vs. random ads, just ads on/ads off)

Consumer Surveillance and Financial Fraud | NBER showed that Apple App Tracking Transparency, which reduces the ability of apps to personalize ads, tends to reduce fraud complaints to the FTC. We estimate that the reduction in tracking reduces money lost in all complaints by 4.7% and money lost reported in internet and data security complaints by 40.1%.

Balancing User Privacy and Personalization by Malika Korganbekova and Cole Zuber. This study simulated the effects of a privacy feature by truncating browsing history for some Wayfair shoppers, and found that people who were assigned to the personalized group and chose a product personalized to them were 10% less likely to return it than people in the non-personalized group.

The Welfare Effects of Ad Blocking by Lin et al. was different—members of the treatment group got an ad blocker affecting all sites, not just one retail site.

[P]articipants that were asked to install an ad-blocker become less likely to regret recent purchases, while participants that were asked to uninstall their ad-blocker report lower levels of satisfaction with their recent purchases.

The ad blockers used in that study, however, were multi-purpose ones such as uBlock Origin that block ads in general, not just personalization.

The effect of privacy settings on scams goes two ways: you can avoid being specifically targeted for a scam, but you can also get more scam ads by default if you feed in too little info to be targeted for the good ads.

The Internet as a whole is much more various in seller honesty level than the Wayfair platform is, which might help explain the difference in customer satisfaction seen between the Korganbekova and Zuber paper and the Lin et al. paper. Lin et al. showed that people were more satisfied as customers when receiving fewer ads in total, but they might have been even less satisfied if they received more of the lower-quality ads that you’re more likely to get if adtech firms don’t have enough data to target you for a bigger-budget campaign.

Another related paper is Behavioral advertising and consumer welfare: An empirical investigation.

The presence of low quality vendors, along with the recent increase in the use of ad blockers, makes it increasingly difficult for new, high quality vendors, to reach new clients. Consumers benefit from having access to new sellers that are able to meet their needs through behavioral ads, as long as they are good sellers.

but

targeted ads are more likely to be associated with lower quality vendors, and higher prices for identical products, compared to competing alternatives found in organic search results

If you look back on the history of advertising, there has never been an ad medium that required so much legal and technical complexity to try to get people to accept it. Why is Meta going to so much trouble to try to come up with a legal way to require people in the EU to accept personalized ads? If ad personalization is so good for consumers, won’t they pick it on their own? Anyway, I’m looking for research on how personalization and privacy choices affect customer satisfaction.

More: building wealth the privacy way?