We use cookies. You have options. Cookies help us keep the site running smoothly and inform some of our advertising, but if you’d like to make adjustments, you can visit our Cookie Notice page for more information.
We’d like to use cookies on your device. Cookies help us keep the site running smoothly and inform some of our advertising, but how we use them is entirely up to you. Accept our recommended settings or customise them to your wishes.

7 Algorithmic Models Used to Manage Personalization Recommendations

Did I miss one?

I think I got them all. I am not 100% certain, but I think I did.  

But let me stop for a second, take a breath, and explain what I am trying to validate.

“And breathe.”

Marketing has started leveraging automated algorithmic models, as part of their decision orchestration platforms, to manage, interact, and personalize customer experiences. This is in addition to rules-based, triggers, A/B testing, multivariate, and segment targeting. The difference here lies in the term “automated algorithmic models.” 

“Does it seem like I am trying to get a couple of keywords in for search?”

Decision orchestration platforms (DOP) are becoming a standard part of the marketing technology landscape. Some of the primary differences between a DOP and campaign management platforms (CMP) is that a CMP operates at a campaign execution level, while a DOP manages real-time interaction events. For a company to operate at real-time interaction events level, they have to rely on automated (sometime predictive and self-learning) algorithmic models.  

“And yet another breath, as that was almost an incisive rant.”

To that end, I tend to utilize and deploy about seven (foundational) types of automated algorithmic models. They include:
  1. Item Grading — If you bought one item, you will also need the following items. For example, if you bought book one of a series, you will need book two and three of the series.
  2. Attribute Filtering — This leverages various attributes about a previous purchase to recommend other purchases. For example: you liked Jurassic World (action movie) with Chris Pratt (actor); therefore, you might like Guardians of the Galaxy.
  3. Product Affinity Modeling – If you like one product(s), you are likely to like these related products. For example: you bought Batman and Superman graphic novels; therefore, you might like Justice League: Gods and Monsters. (Note: shameless fandom plug.)
  4. User Affinity Modeling — Leveraged quite often by Amazon. For example:  you bought a product. Other people who bought that product also bought (or browsed) these products.
  5. Social Media Relation — Leverages social derived data, related to the consumer, to make a recommendation that “your friends are watching YouTuber Tyler Oakley so you might be interested in following YouTuber x.”
  6. Next Best Offer — This approach looks at recent consumer attributes and interactions, compares those attributes and interactions to other individuals, and then calculates the likelihood of the consumer to respond.  It uses this calculation to determine the next best offer to present to the consumer. 
  7. Model Defined — This approach leverages models developed to target and personalize on key components, e.g., brand or behavioral models. Some of these models are built and deployed by your nerdy data scientist, while others are inherent to the solution of choice.

While I think that I got them all, I’m not 100% certain, so I am asking you. Did I miss one? Did I cover the majority?  

Did I even make any sense whatsoever?

The great thing about all of these models is that it moves companies away from the “batch-and-blast” campaign approach to personalized targeted offers. These models enable dynamic personalized offers to the consumer as they move throughout a website — or even as they scroll down.