Login / Register
Let’s say you like to grill — so much so that you spring for a new top-of-the-line BBQ grill. Then, only moments after your online purchase, you’re inundated with ads for the very same grill.
And it doesn’t stop. Two weeks after your purchase, it seems every other ad is for the very grill sitting on your back patio. What’s up? Isn’t personalization supposed to figure this stuff out?
It’s happening now.
Recommendation engines
Recommendation engines have been the cutting edge of marketing for the last 10 years, dramatically powering e-commerce. And no wonder: Harvard Business Review points out that personalized user experience can deliver five to eight times the return on investment on marketing spend. Toward Data Science notes that, at Netflix, over 80% of stream time is driven by personalized recommendation. It’s been working well.
Historically, these engines use a combination of filters:
Generic, which identifies items that are similar to what a user searched for or what is most popular.
Content, which examines user history, identifies keywords and makes suggestions of similar content.
Collaborative, where a user is assigned to a group, based on that user’s history, and items liked by other members of the group are presented.
Ensemble, which is a combination of the above.
While recommendation engines have been highly effective for retailers, there are shortcomings for other areas of business, like the barbeque grill problem for advertisers.
Associative AI
Enter associative AI, where a model is trained to understand that you’re interested in the grill, that you have purchased the grill, that you are a certain type of shopper, and that you are likely to need grill-related items.
It also takes into account grill product inventory and product discounts across retailers. In fact, the parameters here are almost limitless. The more data the model considers, the better its suggestions will be. This complexity is why simple recommendation engines have difficulty with relevancy and why associative AI is critical.
As Oxford University notes, “New AI uses associative learning techniques rather than AI’s traditional neural networks to challenge the conventional wisdom that artificial neurons and synapses are the sole building blocks of AI.”
With associative AI, you are no longer inundated by ads for the same item after you buy your grill. Instead, you’ll see ads for complementary products: a BBQ cookbook, an instant-read thermometer, a cleaning brush or Kansas City steaks.
Generative AI
Sophisticated associative AI recommendations are just the beginning. Next comes generative AI, which could make the purchase relationship much more effective and efficient for the retailer and much more relevant and efficient for the consumer. If basic personalization can deliver five to eight times the ROI on marketing spend, imagine what generative AI can do when ads are personalized beyond “insert customer’s name here.”
For the retailer, generative AI makes ads smarter. While associative AI is guiding the user through the purchase process, the actual ad may be generated by AI, so that every time a user engages, the advertiser is offering a highly personalized, intelligent conversation regarding grill care, techniques and additional products of interest. This “conversation” and any associated products become highly relevant to the user.
The ads would be significantly more effective since generative AI hones the message, creates a unique ad, and determines the format required for where the user will be viewing it. This becomes an enormous advantage for retailers to develop significantly more relevant, effective marketing.
In addition, because of AI’s sophisticated targeting, advertising dollars are not spent on the wrong people. A business is targeting the right audience and only them, saving time and money. Customer acquisition costs plummet.
Consumer experience
This combined use of AI can also enable a better cross-platform user experience. Does eBay know you bought a grill on Amazon? No, and because of data privacy, this information sharing likely won’t happen anytime soon. But technology like a personal AI assistant built into your device knows and could tie it all together.
Now you essentially have a personal assistant that delivers highly targeted, highly relevant recommendations based on a thorough understanding of who you are. For instance, AI understands when you bought your grill and could message you when it’s time for new pellets or a deep cleaning. Generative AI becomes seamlessly migrated into your everyday life, acting as your crystal ball, with everything based on you.
Yet, for generative AI to grow, it must be able to support e-commerce recommendations based on billions of parameters — all at lightning speed. To deliver on this need, Micron has developed an unparalleled memory and storage portfolio, uniquely positioned to fulfill the demands of AI — from data centers to devices. Our products can help engineer what generative AI will be capable of next. With this type of performance, AI is better for businesses and better for consumers.
So technology really can enrich life for all. On that note, let’s fire up the grill!