Search

WordPress AI Recommendations
- SEO, conversions -

Table of contents :

Algolia recommendations guide

Algolia logo : search & recommendations provider

Table of contents :

Algolia is one of the leading providers of search and recommendation solutions, with a significant market share in the industry. So considering using Algolia to power your website’s search & recommendations is a no brainer.

 

Why use AI recommendations ?

 

AI recommendation models are trained on user interactions or clicks to recommend the most relevant products or items for each user.

 

You can usually find them in e-commerce websites, at the bottom of products detailed view or in the homepage. But they can also be found on streaming websites (movie and music) or services, news and many other industries.

 

Algolia AI recommendations in Gymshark

Algolia AI recommendations in Gymshark

 

Recommendations are useful since they can increase user engagement and drive conversions. Since users are continously served the content or items they want to see, they aremore engaged with your platform, spending more time exploring and interacting with recommended content. Additionally, by presenting users with personalized suggestions that align with their interests and preferences, recommendations can effectively guide them towards making a purchase or taking a desired action, ultimately driving conversions and maximizing revenue.

 

Algolia recommendations features

 

Algolia offers multiple recommendation models for all purposes. They are based on content-based and collaborative filtering algorithms.

 

Recommendation types (models)

 

Related products

Image algolia-model-related-products.png of Algolia recommendations guide

The “related products” recommendations can be used to recommend items or products similar to the ones the visitor is currently viewing or has viewed in the past. This will offer the visitor a more diverse catalogue since they can browse countless other similar items that could interest them.

 

Frequently bought together

Image algolia-model-frequently-bought-together.png of Algolia recommendations guide

The “frequently bought together” recommendations will, as it’s name suggests, recommend items that are not similar but could complement eachother and are frequently bought together.

This means that when viewing a hat, this recommendation will probably suggest sunglasses, sunscreen and shorts.

 

Trending items

Image algolia-model-trending-items.png of Algolia recommendations guide

Algolia AI recommendations (trending) in Gymshark

Algolia AI recommendations (trending) in Gymshark

 

The “trending” recommendations will recommend the most popular items to the users. This can be displayed on our homepage so that your users can know where to start.

 

Looking similar

Image algolia-model-looking-similar.png of Algolia recommendations guide

Recommend visually similar items using this top of the line AI recommendation. The AI model will recognize items or products that have similar looking feaured images and recommend the relevant ones.

 

Recombee also provides a visually similar recommendations and we have a demo available for you to check out the results.

 

Trending facet (filters) recommendations

Image algolia-model-trending-facet-values.png of Algolia recommendations guide

The trending facets recommendations will recommend facets that have recently gained popularity.

 

Recommend rules

 

Set the Algolia recommend rules

 

You can further customize your recommendations without having to modify your code using the recommend rules that are applied in the Algolia backend.

 

The rule in the example above will only display items for which the price is greater than 10.

 

This is offers the users total freedom. The screenshot above shows a rule created visually in the Algolia dashboard but you could also create them using JSON.

 

Algolia UI recommendations libraries (Javascript & React)

 

Algolia offers React & Javascript UI libraries to easily display your recommendations to your visitors in the frontend. Algolia provides an example script so you can instantly set up your recommendations :

/** @jsx h */
import { h } from 'preact';
import {
  frequentlyBoughtTogether,
  relatedProducts,
} from '@algolia/recommend-js';
import recommend from '@algolia/recommend';

const recommendClient = recommend('YourApplicationID', 'YourSearchOnlyAPIKey');
const indexName = 'YOUR_INDEX_NAME';
const currentObjectID = 'YOUR_OBJECT_ID';

frequentlyBoughtTogether({
  container: '#frequentlyBoughtTogether',
  recommendClient,
  indexName,
  objectIDs: [currentObjectID],
  itemComponent({ item }) {
    return (
      <pre>
        <code>{JSON.stringify(item)}</code>
      </pre>
    );
  },
});

relatedProducts({
  container: '#relatedProducts',
  recommendClient,
  indexName,
  objectIDs: [currentObjectID],
  itemComponent({ item }) {
    return (
      <pre>
        <code>{JSON.stringify(item)}</code>
      </pre>
    );
  },
});

 

Collect user events

 

Events are actions that users take on the website. They are used for all types of personalization since they allow the AI models to learn more about each one of the users. This means that as user engagement increases on the website, the accuracy of recommendations also improves.

Image wpsolr-enterprise-algolia-account-events-detail.png of Algolia recommendations guide
wpsolr enterprise algolia account events detail

 

Image wpsolr-enterprise-algolia-account-events.png of Algolia recommendations guide
wpsolr enterprise algolia account events

 

Algolia provides many user events for you to collect :

  • clickedObjectIDsAfterSearch
  • convertedObjectIDsAfterSearch
  • addedToCartObjectIDsAfterSearch
  • purchasedObjectIDsAfterSearch
  • clickedObjectIDs
  • convertedObjectIDs
  • addedToCartObjectIDs
  • purchasedObjectIDs

 

Related posts ... not powered by WPSOLR 😊

Release 22.3 with Weaviate vector search

This is an exciting moment, where #vectorsearch is getting to the rich #wordpress and #woocommerce community: site and shop owners, agencies, or even specialised hosting