When re-targeting works

by Nigel Hollis | February 25, 2014

Apparently, I am not the only one who dislikes the practice of re-targeting. Bill Pink passed on this tweet from Sarah Silverman, “Amazon I love u but just b/c I bought a trash can doesn't mean i need all ur latest trash can news. I got one. I'm set.” 

Of course, she is right. Amazon ought to know she has bought a trash can, and maybe it does, but the simplistic mindset that behind these systems works on the principle of, “You looked at this so you must be interested.” Actually, no, for most things that we buy in life, we take them for granted and don’t think about them again until we need to replace them. Dynamic retargeting, where the ads feature a specific product that the consumer has previously looked at, seems all the more annoying if you now own the item being advertised.

I think that the problem underlying the use of re-targeting is the same one that underlies most marketers’ use of digital advertising and social media. Marketers assume that people are far more interested in their brands than they really are. For instance, most people do not want to hear from the brands they buy, let alone have a conversation with them. Instead, people want brands to do the job for which they were acquired and to shut up. 

Failing to respect the interactive nature of digital media is not without its risks. When using dynamic retargeting, marketers are applying a sales pitch mindset to a medium where the recipient can cut off messaging at the click of a button or use the same medium to fight back. Sarah Silverman may not have paid to promote her tweet, but she could easily have done so. Hasan Syed, promoted his tweet in New York City and in the U.K. when British Airways lost his luggage, stating, “I refuse to stop running Twitter Ads until @British_Airways finds the lost luggage.” 

In reply to my previous post about the merits of re-targeting, John Were did share some results that indicated re-targeting was effective (under the tested circumstances). 

Brand Premium is the new title of Nigel Hollis' Book

In googling the topic, however, I turned up a paper by Anja Lambrecht and Catherine Tucker titled “When Does Retargeting Work?” that finds that dynamic retargeted ads are on average less effective than their generic equivalent (again, within the parameters of that specific test). Dynamic retargeted ads only proved more effective when consumer’s behavior indicated a continued interest in making a purchase, for instance, after they had visited an online review site. 

That makes perfect sense to me. If someone looks at a product and then uses a review web site it is a strong indication that they have not yet made up their mind about which brand to buy. A timely reminder might swing their decision in favor of the advertised brand. Otherwise there is a good probability that, like Sarah Silverman, the recipient has already made their choice and the exposure is wasted. 

Putting a little more thought into how and why someone is retargeted, could go a long way to save money and boost effectiveness. So what do you think? Please share your thoughts.


Leave a comment
  1. John Were, June 23, 2014

    Thanks for the reference Nigel.

    To comment on this post specifically, it is pretty silly (and mostly avoidable) to show already purchased products in ads immediately after the purchase if the product is something like toothpaste; but if it is toothpaste it might make sense to show the same product 30 days later when a repeat purchase is likely. In the meantime the ads can display best-selling products, products with the biggest margin, seasonal products, clearance products and so on. It's a dynamic catalogue which, if powered by the correct data and logic, can increase purchase behaviour.

    Messaging which doesn't involve products is great too and can be used to announce discounts, re-branding, new product lines and all the other kinds of messages. This kind of messaging can be interspersed with product-driven ads with algorithms driving which ad to show when in order to maximise results.

    The trouble is, it's easy to game the attribution of retargeting. Many companies have built large businesses based on it. These businesses regularly arbitrage ad cost into a 'result' and generate massive margins by doing so. These companies aren't always interested in the verifiable effect, they are interested in their sales pitch to the advertiser and their own bottom line.

    Ultimately those seeing the ads are customers, or people with some kind of interest in becoming a customer, or people with friends who might become customers and as soon as they've visited your website you've got an opportunity to talk to them for the lifetime of the cookie on their browser. This opportunity to talk can't be missed but it can be planned and executed carefully with appropriate monitoring of results.

  2. Claudia , June 23, 2014

    There is a lot of truth in these words indeed. I believe that re-targeting can work for some industries: Women shop around for fashion items without necessarily making a purchase, just to get some inspiration. So why not remind them of the items they have viewed before?

    Still, overall this practise can be perceived as so annoying that it may even harm a brand. I wonder whether it would be more effective to use simple branded banners for re-targeting, i.e. to put a brand back into consumers' minds instead of the product viewed?

  3. Mike Connolly, June 23, 2014
    Agree, this is good sense and bears testing.  

    Leave a comment