Step by step guide to implement simple Recommendation Engine using Google Optimize
One of our e-commerce client (OFLARA) who sells high-quality crystal jewelry online approached us with an idea to implement a product recommendation section to upsell on Add to Cart popup. This recommendation section enabled their customer to easily add more products to cart; therefore helped increase revenue for our client.
Below is the step-by-step guide on how we implemented this recommendation section.
Step 1: Getting the product catalog (Excel, CSV)
We received an excel file containing the list of recommended products. We opened the excel file as a CSV file, so we can easily convert the data to a JSON object which can be easily returned over the API call.
Step 2: Convert CSV to JSON
To convert the CSV into JSON we used http://www.convertcsv.com/csv-to-json.htm
Now we can call the API and get the recommendation results in JSON format. We then display the recommended products on the pop-up overlay. We created the API using Amazon Serverless. Below are the steps on how we implemented Serverless:
Below is the function to create the endpoint which processes the request and returns the results in JSON format.
The website is built on Shopify so we made sure to use Shopify’s API to send Add to Cart request.
We completed this test within 10 hrs including QA and bugfixes. Hope you will find it useful if you want to implement a similar recommendation section on your website.
We ran the test for 30 days and we noticed that there was a significant improvement in the overall revenue. Below are the result screens
After seeing the result the client was happy and made us run the test to show to 100% of the traffic. This test is still live and if want to see how its working you can follow the below steps:
Go to www.oflara.com
Hover on any jewelry item and you will see “Add to Cart” button just click on that which will open a pop-up like the below:
Now, on Add to Cart pop-up we can see the products under “You might also like” heading