Optimize Amazon's Ad Retargeting

Optimize Amazon's Ad Retargeting

Everyone is doing online retargeting. However, I am always surprised that most fail to exclude customers who have already purchased the item.


"Retargeting" is the advertisement technique that allows companies to follow you around the web. After viewing a product on their website, they can "retarget" the ad systems of Facebook, Instagram, and Google, to display that product on social media and other sites. These campaigns are easy to set up. Retargeting works to increase sales, especially when used early - within hours to days.

Stop retargeting users who just purchased

Showing adds to a customer for a product they just purchased does not make sense. However, I see this happening all the time. Companies of all sizes make this mistake.

This happens I suspect either because these organizations haven't set up their retargeting campaigns properly, there is an error in the Audience selection functions of Facebook, or their refresh times on their campaigns is too slow to account for the fact a purchase has been made.

In the example below, is promoting a book I already purchased. This is a waste of advertisement money for them, I've already bought it. It also creates friction between the customer (who already purchased the item), and the business who seems at this point overbearing to continue to show adds for something already purchased.

Customers who have purchased an item, should be excluded from that the retargeting campaign. Facebook demonstrates a simple way to exclude customers who purchased here. For automated retargeting campaigns for specific products the work to exclude customers is more complex, but definitely doable. Especially as the same email address and tokens are used across the platforms.


Simple version of Facebook retargeting exclusion list


Retargeting the right way

Targeted ads should be used to sell products businesses anticipate customers want. In the case of books, as a reader completes a book they enjoyed it makes sense to push ads for the next book in the series, the latest book by the same author, or another strong book in the field.

For a company such as Amazon who owns both Kindle and Audible they have the metrics needed to know how far a reader is in the book and infer if the book was a hit.

Understanding a user's review of a book can be determined after they complete the book by asking them to rate it.[1]  A better way is to infer their rating of the book prior to them completing it - so that ads can start prior to the book's completion.

The reader's satisfaction of the text can be inferred based on how they consumed the book vs other books in their library. Metrics can be compared such as:
- how quickly the book was started relative to its purchase and download
- how quickly it was completed
- if it was read cover-to-cover, or with other books at the same time
- if it was read at 'extreme' hours of the day
- number of of re-listens, rewinds,
- number of notes and markup
- number of other books on the subject and by the author in a user's collection

Being able to quantify a user's satisfaction with a product, helps determine how aggressive the ad campaign should be. If the user really liked the book the ad campaign should be more aggressive in anticipating a sales conversion and determining ad spend and discounts.

Ideally, as I near competing a book I liked, retargeted ads should offer discounts for the next book by the author. The discounts should be tied to a narrow time window to add time pressure and correspond with having the book and the value I obtained from it front of mind.  

In the example above of Blue Ocean: Strategy, the book Amazon was advertising to me after I purchased it already, the metrics would show I enjoyed the book and would be interested to read another by the author. The book has an obvious sequel (In fact, I even looked at the book on Amazon/Audible prior to purchasing this book). However, Amazon/Aduble has not sent any online retargeted ads, emails, or in app adds encouraging purchase of the next book. The only time Amazon has so far offered another book by this author was the single screen shown once the book is competed where it asks you to rate the book and shows you 'other books like this'.



A week later, Amazon is still retargeting adds for the same book I've already purchased.

They also are retargeting adds for items that are 'single purchase categories'. For instance, several weeks ago I bought on Amazon a travel neck pillow. I consider this a single purchase category - once one item in the category has been bought, the customer is not inclined to purchase a second.  

Items in single purchase categories should be excluded from retargeting campaigns.  

Unless, the customer returns the original item dissatisfied. In that case, spam them with retargeted adds from that product category. Particularly items with high user user satisfaction ratings.


[1] At the end of each book Audible asks the user to rate it. Those ratings 'appear online'. The fact those ratings appear publicly online are a deterrent to me rating the book. I suspect others feel the same way. I'd like to rate all the books in my Audible library in the hope Audible's recommendations would be more accurate. However, will only do so if the ability to share the ratings publicly can be disabled. - cost & financial data visualizations - cost & financial data visualizations

Disinfection wipe containers require restraining brackets

Disinfection wipe containers require restraining brackets