According to researchers at Stanford, highly targeted ads may not be all they're cracked up to be.

Based on a mathematical model they built based on game theory, the researchers instead suggest that advertisers "prefer to remain in a state of partial willful ignorance so as to preserve communication credibility."

Eilene Zimmerman explains...

In this case, the researchers were looking at cheap talk in retail, for example, an ad promising 'Lowest Prices in Town'.

That can be credible when it’s used to draw in appropriate customers; in this case, those who are price sensitive.

At the same time...

They found that the most personalized ads were less effective because consumers worried they were being exploited.

For example, says [Stanford Graduate School of Business professor Pedro Gardete], someone looking for a prom dress 'might get an ad from a retailer saying, "We have a wide selection of prom dresses! Click on this link!" The consumer clicks, and it turns out the retailer has dresses for all occasions but not specifically proms,' says Gardete.

Those kinds of ads frustrate consumers and eventually become meaningless to them.

Based on this, Gardete suggests that businesses might adopt a "less is more" approach in which less information is collected, information collection is more transparent, and targeting is used more sparingly. 

Theory versus reality

While there's no doubt that a growing number of consumers are concerned about their privacy and how marketers are using information to track and target them, given the continued level of interest and investment in targeting tech and targeted ad offerings, does the researchers' model actually reflect reality?

Obviously, a hypothetical retailer falsely promoting that it has a wide selection of prom dresses when it doesn't isn't likely to see good results, but this isn't how most experienced digital marketers are operating.

Instead, retargeting (and personalisation) are widely seen to drive ROI in the real world.

As an example, AdRoll, a performance marketing platform provider, detailed in a case study (PDF) how one apparel retailer used retargeting to deliver a 10.5x average ROI, 13% conversion lift and 33% lower CPA than average for other apparel retailers.

Facebook Custom and Lookalike Audiences have delivered similarly impressive results.

Crowdfunding platform Tilt doubled its conversion rate using Custom Audiences, and lowered its mobile cost per install by 30% using Lookalike Audiences.

And Hospitality giant MGM realized a greater than 5x return on spend using Custom Audiences.

Needless to say, any specific marketer's mileage will vary, but on the whole, marketers are becoming more and more adept at targeting consumers online and doing so to good effect. 

That doesn't mean that marketers should rely on targeted ads exclusively, and the Stanford research is a reminder that targeted ads need to deliver what they promise to consumers.

But targeted ads are here to stay because they work well enough of the time, even if many consumers say they don't like them.

Patricio Robles

Published 23 May, 2016 by Patricio Robles

Patricio Robles is a tech reporter at Econsultancy. Follow him on Twitter.

2647 more posts from this author

You might be interested in

Comments (5)

Pete Austin

Pete Austin, Founder and Author at Fresh Relevance

This research supports my experience over the years that personalization needs a light touch, but the conclusions go to far. Adopting 'a "less is more" approach in which less information is collected' is not a good idea - more data is always better than the alternative.

Problems like the example could occur when you have very little data. For example if you have a single data point for a shopper - that they once ordered a prom dress - you might target narrowly based on that one product SCU, with unfortunate results.

Good marketers wouldn't do this, but can configure recommendations that pick a wider range of products, based on prom dresses having categories like "young", "female", "dress" and maybe price range and dress size, as well as the SCU. And most importantly marketers would also include unrelated recommendations for unrelated popular clothes lines, using crowdsourced data.

about 2 years ago


Sebastiaan Elsholz, Consultant at Newcraft

Thanks for posting this article. It raises an interesting question: is over-personalization counter productive? However I’m not sure the example given (prom dresses) proves this point. It could also be there was a misfit between the ad and the landing page. I can imagine a highly personalized ad leading to a much less or no personalized page would frustrate a consumer. But the solution there could also be to make sure the landing page is personalized much more, just like the initial ad. In other words: doing the exact opposite of what the article suggests.

about 2 years ago

Chris Perry

Chris Perry, Founder at No-Brainer Marketing

Driving a unique set of leads to a generic page/site is a structural failure of the overall campaign strategy, not the ads targeting.

Rookie mistake that would have been avoided by consulting with a digital advertising professional prior to testing.

The ads worked...that's the easy part.

about 2 years ago

Simone Kurtzke

Simone Kurtzke, Lecturer in Digital Marketing at Robert Gordon University

Agree with Pete, the issue is most likely not enough (or not the right) data to target effectively. I for instance am not really targetable in the narrow sense (intentionally so) even though I am online a LOT (and have been for 20 years now).

There are also a ton of users who don't understand retargeting and find it creepy. E.g. once I explain to my students that their personal data, online behavioural data (such as browsing), as well as online relationships now have economic value and that their personal data is mostly sold to advertisers, some of them are shocked as they were not aware of this (or at least not of how far and how deep this goes).

People who don’t work in digital (this includes Stanford researchers!) can only understand this from their outsiders’ perspective. Some of my students are avid users of social but have little comprehension of how it all works (that THEY are the products) – they think it’s all about sharing and being friends. Many especially younger students are also vulnerable to intimacy creep as practiced by the major social networks, and really quite easy to manipulate as a result.

I'm interested in doing some research around that in the future as I'd like to explore it in more depth. I've also observed that young people appear to be more at ease with transactional-type relationships with brands and people so there's a blurring of boundaries, which may also explain why e.g. influencer marketing is so big (they don't mind being soft sold stuff by influencers, so this whole 'native advertising' discussion perhaps also needs a rethink).

Certainly I’m watching all of this with great professional interest, including related issues such as adblocking etc., although in terms of personal interest the ONLY thing I'm excited about these days is VR (bring it on PS4!).

about 2 years ago

Parry Malm

Parry Malm, CEO at Phrasee Ltd.

Reminds me of the time my friend got a perfectly personalised valentines email for a toilet brush. Nothing says "love" like a clean bog. (She didn't buy it, btw).

about 2 years ago

Save or Cancel

Enjoying this article?

Get more just like this, delivered to your inbox.

Keep up to date with the latest analysis, inspiration and learning from the Econsultancy blog with our free Digital Pulse newsletter. You will receive a hand-picked digest of the latest and greatest articles, as well as snippets of new market data, best practice guides and trends research.