Do you want to know what your users are doing as they search and browse your online catalog? Do you want to suggest recommended products as they shop? 

If you are not yet capturing your customer behavior, and not using Recommend – Turn it on today!

We’ve all seen and clicked on recommendations while shopping online. As soon as we add an item to the cart, we are prompted to continue shopping for items Frequently Bought Together.  

Take a closer look. Those recommendations may have been powered by Algolia Recommend – which in turn is powered by Algolia’s out-of-the-box analytics. 

By capturing user behavior with Algolia Analytics, and feeding that data into Algolia Recommend, you can start offering meaningful recommendations to your customers that will increase their satisfaction and conversion rates.

Consumers appreciate it when an online store suggests relevant items in a non-invasive and timely manner. Consider Jennifer, who is looking for a bed for her sick dog, Luna. She finds the perfect bed and adds it to the cart. Then she notices a “Frequently Bought Together” section that recommends dog covers, pillows, and other accessories that could comfort a sick dog. Jennifer may also see items that other dog owners have bought for their beloved pets. 

This is how Algolia Recommend increases your customers’ basket sizes and values while improving their experience and loyalty.

But how does Algolia know what to recommend? Does Algolia know that Jennifer is a pet owner or that Luna is sick?

The simple answer is no. Algolia doesn’t know anything about Jennifer except that she is following a consumer pattern based on similar purchasing behavior by similar customers. Many customers like Jennifer have searched for, clicked on, and bought similar products.

And while it may sound simple, to mix and match what customers want, it actually takes a lot of analytics data and a careful implementation of AI-driven machine learning to get it right. 

Recommend works by collecting user behavior (event-based analytics) and training  models to display features such as: Frequently Bought Together, Trending Items, and Those Who Bought This Also Bought. These models have been carefully designed to align with a customer’s original purchasing intent, so as to guarantee a greater likelihood that the customer will continue shopping. As you can imagine, only the most relevant recommendations drive financial growth and customer loyalty.

With Algolia Recommend you can significantly improve your business performance

Before telling you how to set it up, let’s see how the right recommendations can lead to more upsells, cross-sells, and overall customer loyalty (Ask Amazon!)

Gymshark has boosted its revenue by adding Algolia Recommend to its Algolia Search implementation. Gymshark’s success metrics:

• 150% increase in order rate and 32% “add to cart” rate with new users on Black Friday
• 13% higher order rate and 10% higher “add to cart” rate from returning customers
• 1.4 clicks per user vs. 1.1 with previous solution

Gymshark adds Algolia Recommend to handle the crucial Black Friday period. We’ve discovered that training your recommendations models during big shopping events such as Black Friday or Christmas Holidays will help you understand your shopper’s behaviors in time for the next big event.  

Setting up recommendations – All you need to do is start training the model!

Here are the four steps to Recommend.

1. Capture user activity (events)
2. Send the events data to Algolia
3. Enable Algolia Recommend on the Dashboard and have it start generating recommendations in the background
4. Display recommendations to your users

Step 1 – Capturing quality data

Data quality is essential. Steps 1 & 2 are entirely in the control of the customer. Customers build up their analytics dataset by capturing significant user activities (events). This means that you must be careful to capture actions that are “significant” to your business. Of course, purchase history is almost always relevant, but what about the shopping cart? Or category pages? Facets? 

Essentially, you want to track all user actions that you believe will have a significant likelihood to lead to a conversion. 

The good news is that Algolia comes with an Analytics Dashboard that you can use to test the quality of your data.

Step 2 – Sending events to Algolia

How do you send Recommend events? Like most analytics, your engineers need to add a line of code to the pages in your website where you want user events to be captured. For example, on the search results page, you’ll add a line of code to collect every click on a result and send it to Algolia. On the purchase page, or the view product page, you’ll add a line of code to capture the product just purchased or viewed, and then send it to Algolia. 

In sum, whenever you want to capture an event, you need to (1) add a line of code that captures the information that your user is doing and (2) send it to Algolia via the Algolia’s API. One line of code per event.

Steps 3 & 4 – Enable recommendations, and you’re good to go

Once you’re sending significant user actions (here’s a big list), then all you need to do is enable recommendations in Algolia’s Dashboard (step 3) and let the machine roll. As you send events, Algolia does the rest – it feeds those events into its machine learning models and generates a robust set of recommendations. 

After a short time – could be weeks, depending on the activity of your website –  you can start displaying those recommendations to your users (step 4). Here are some powerful ways to leverage recommendations, and an overview on how to display recommendations.

The bottom line

If you can provide great recommendations, ensuring that your customer journey is enjoyable and stress free, you’ll be on your way to building customer loyalty that will transform your business bottom line.

Our powerful Recommend API and AI-driven technology will quickly and seamlessly improve your user’s experience and the conversion rates on your website and mobile applications.

To learn how leading ecommerce retailers leverage the capabilities of an AI-powered recommendations engine to exceed and transform their digital merchandising goals, contact us, and let’s get rolling!