Will Google Recommendations AI beat your product recommendation model?

YouTube, Google Search and Play use one of the largest and most sophisticated industrial recommendation systems ever. It didn’t take long for Google to make this knowledge and computing power available outside of their products. Today I would like to highlight briefly Google’s brand-new packaged AI product: https://cloud.google.com/recommendations/.

The out of the box fully managed Google Recommendation AI service is currently in closed beta. Yet, Crystalloids has been whitelisted as the only Google services partner in the Netherlands to test and implement it for eligible clients.

We are currently working on the implementation of Recommendation AI at a couple of leading e-commerce brands. The results will be disclosed at the end of December.

WHY might Google Recommendation AI beat your model?

1. Delivers highly personalized recommendations at scale
2. Leverages the best algorithms in the market
3. Adapts in real-time to user behaviour
4. Fully managed service
5. Shocking good results in the USA already 

What’s more, you don’t have to be on Google Cloud Platform with your products and user data to start using Recommendation AI. Just bring the required data to GCP and you will get responses back to connect to your paid and owned channels by your platform of choice.

HOW does Recommendation AI work?

Three steps to take:

A Ingest your data       B Customize your recommendations      C Deliver to channels

A Ingest your data

Start ingesting the products from your catalogue that you want to use. Of course, you will bring your customer data in as well. The architecture might look like this when the recommendations are pushed to the website:

B Select what you want to achieve

Select recommendation type

• Recommended for you e.g. on home page
• Other items you may like e.g. on the product detail page
• Frequently bought together e.g. in shopping cart view/ checkout start
• Recently viewed not a recommendation but useful on e.g. home page

Select objective

• Click-thru rate emphasizes engagement / maximize the likelihood that the user interacts with the recommendation
• Conversion rate default optimization objective for the "Frequently bought together"
• Revenue per session maximizes the likelihood that the user purchases the recommended item

Set business rule

• Filter on articles that are out of stock
• Filter on duplicate items
• Filter by a custom tag
• Result diversification on/off
• Personalization on/off

C Deliver to channels

You can deliver to any channel such as Web, Mobile web, Mobile apps, email, social media, etc

WHAT needs to be done to test Recommendation AI?

• Define success criteria for Recommendations AI
• Build pipeline for transferring product catalogue data to Recommendations AI
• Build pipeline for transferring user-events data to recommendations AI in real-time via Pub/Sub and Dataflow
• Build pipeline for backfilling historic user-events from BigQuery to Recommendations AI
• Customize Recommendations AI for delivering personalized content on a website channel
• Test Google Recommendations AI model for delivering personalized content on the Web, Mobile web, Mobile apps, email, social media or any other channel of your choice