How Algolia works

Algolia is a hosted search engine capable of delivering real-time results from the first keystroke. Algolia’s powerful API lets you seamlessly implement search within your websites, mobile, and voice applications. The Algolia search engine powers billions of queries for thousands of companies every month, delivering relevant results in under 100 ms anywhere in the world.

The first step is to send your data to Algolia, in a proper format. Once on Algolia’s servers, you can start testing search with your data. You can do this without code, using the dashboard, or with code, using one of the official API clients. At this stage, you’ve got the basic structure of an Algolia search solution.

There’s much more to creating a full search implementation, and the documentation walks you through every possibility.

The Algolia engine

Speed

The Algolia search engine was built to ensure speed at every point in the processing workflow. For example, search requests always have priority over indexing operations, to guarantee an optimal experience for your end users.

Relevance

The main goal of the Algolia engine is to find all records that match a query, then to order them with the best ones first. You can find more about what relevance and ranking mean in the documentation.The overarching principle is that Algolia uses a tie-breaking algorithm with a variety of criteria to weigh and compare matching items against each other. This ensures that the best matches appear at the top.

Here are some examples of the ways the engine compares records.

• Is there a typo? If yes, how many?
• Does the query match on the first letter? Does it match the whole word? Or only partially?
• Does the word match a synonym?
• Does the query match the title of a record or its description?
• Is this item more popular than another item?

When you add up the answers to these questions, some records have a stronger or weaker relevance. This is what determines the order of the results.

Transparency

The thoughts and criteria that Algolia uses to find and rank records are available for your scrutiny. Furthermore, you can configure most of it, making all engine defaults adaptable to your unique needs (without secrets or complicated statistical formulas).

An end-to-end search solution

You can’t build high-end search without excellent user experience. This is why Algolia provides you with everything you need to build an as-you-type search experience with instant results. Users greatly benefit from seeing well formatted search results with images and highlighted text instantly on their screen, rather than having to submit their search and wait for results. This is “search at the speed of thought”. Visual immediacy is vital to the decision-making process.

While speed is key to user experience, many other aspects go into making Algolia a full-featured, comprehensive search solution: custom ranking, configuration, and adaptability, tailored UI components for the frontend, personalization, analytics, A/B testing… Algolia provides everything you need to build excellent search.

Implementing Algolia

You can break down how to implement Algolia into a few steps:

• Designing your user experience
• Indexing your data into Algolia
• Building a search user interface
• Configuring relevance and ranking settings

Before starting, it helps to create a plan to go live and make sure your team is ready to go. Once you do go live, you can improve your implementation using Algolia Analytics and A/B testing.

Building a search UI

To help you build the best solution as quickly and with as little friction as possible, Algolia offers a family of UI/UX libraries called InstantSearch. InstantSearch offers a full set of UI components that you can use on any web, mobile, or voice app. In minutes, you get a fully functional search UI. You can use the InstantSearch widgets out-of-the-box, or fully customize them to match your needs.

Insights and analytics

Building excellent search is also about refining it based on how users interact with it. Algolia’s analytics provide constant and reliable feedback. This is crucial to know whether your search solution is achieving your goals. Algolia provides such feedback by capturing user activity and giving you the ability to track click and conversion events, perform A/B testing, and personalize relevance.

Infrastructure

Algolia hosts each subscriber on a three-server architecture called a cluster.

If one or two servers go down, there’s always a third one ready and available. Having all three servers go down at the same time is extremely rare, especially given that they’re not in the same data center. Service availability is one of Algolia’s highest priorities.

Algolia has servers all around the world, allowing you to have your data as close to your user base as possible. It reduces the network distance and latency between a users’ search request and the servers. These servers are highly performance systems, allowing all searchable data to be placed in memory. This plays a significant role in Algolia’s quick response time.

The full ecosystem

To ease the integration Algolia provides:

• A full-featured dashboard to manage data and configurations.
• API clients for many programming languages (PHP, Ruby, JavaScript, Python, Swift, C#, Java, Golang, Scala and Kotlin).
• A set of frontend libraries called InstantSearch to implement Algolia on your website (vanilla JavaScript, React, Vue, Angular) or mobile app (iOS, Android), as well as an Autocomplete library to build autocomplete experiences.
• Integrations for web frameworks, including Laravel, Symfony, Django and Rails.
• Extensions for Magento 2, Shopify and Shopify Plus, Salesforce B2C Commerce, Zendesk, and Netlify.
• An automated crawler to extract content from multiple sites and send it to Algolia.

Further reading