BigAI™ Product Recommendations
Powered by Google Vertex AI Search for retail (opens in a new tab).
Personalized product suggestions based on shoppers' browsing history, purchase history, and product details help shoppers discover relevant products quickly, increasing the likelihood of making a purchase and increasing average order value.
BigCommerce integrates with Google Vertex AI to generate recommendations. You can query product recommendations to display on storefronts using the GraphQL Storefront API after you set up this integration. The GraphQL Storefront API returns product recommendations generated by Google's machine learning (ML) models (opens in a new tab) for a specific shopper at a specific moment in time.
Requirements
- This integration is currently in a closed beta. If you are interested, you can sign up for the waitlist (opens in a new tab).
- This is only available for BigCommerce Enterprise plans.
- You will need to sign up for a Google Cloud account and set up billing to use this solution.
- For now, there’s limited support for headless and Catalyst stores. Find more details in the Stencil, Headless, and Catalyst page.
- You will need a sufficient amount of data to train the ML models effectively. You can find detailed data requirements for each model type and what this integration supports in the Machine Learning Models page.
- We need the data collected from BigCommerce Analytics to train the ML models. Thus, if you've turned off the native BigCommerce Analytics, you must turn it on to utilize BigAI product recommendations. To turn on analytics, navigate to Settings > Security & Privacy in your store's control panel and enable Analytics for my business.
Costs
This solution is run in your Google Cloud account; thus, you will be responsible for all fees from Google. This includes, but is not limited to, prediction requests, model training fees, and search charges. Learn more about Google's pricing (opens in a new tab). BigCommerce currently has no plans to charge additional fees for access to this integration. If that were to change, we will notify you in advance.
Example cost for illustration purposes only
If you have 100,000 views on your PDP (product detail page) in a given month and on your PDP you are showcasing two different product recommendation types (“Others you may like” and “Frequently bought together”), then you will have 200,000 product recommendation requests in a month.
If Google’s pricing is $0.27 per 1,000 product recommendation requests then your product recommendation serving costs will be 100,000 views * 2 models * $.27 / 1,000 = $54. You will also have model training costs. If your model training costs are roughly 30% of your serving costs then your total costs would be $54 + $16 = ~$70 per month.
Overview of how it works
The following highlights the roles that you and BigCommerce have in this integration.
- You create a GCP (Google Cloud Platform) account, create a project in GCP, set up billing in GCP, and make a BigCommerce API call to turn on the integration.
- BigCommerce sends catalog and shopper behavioral events to your GCP account for model training.
- You configure your ML models in your GCP account and then your ML Models will train on the data provided in the step above.
- You get product recommendations from the BigCommerce GraphQL Storefront API and integrate that into your online storefront. When you send a request to the GraphQL Storefront API, BigCommerce sends a request to Google's Vertex AI for product recommendations. BigCommerce receives product IDs from Google, which are used to fetch product information to include in the GraphQL Storefront API response.
- Within GCP, you can get analytics on how well the solution is performing.
Getting started
For more details on how to get started, see the Getting Started guide.