Algolia vs. Elasticsearch

Ease of implementation

For site visitors

The brilliance of consumer-grade search is anticipating user needs. When search and discovery experiences are intuitive, the interface feels like an extension of our mind. A satisfying experience has to:

• put the user in the driver’s seat
• serve instant, search-as-you-type results in your language of choice
• be spelling agnostic, with high typo tolerance
• build a clear, tangible connection between queries and results
• work flawlessly on every platform

This is where Algolia’s AI-enabled features shine brightest. Filters and facets, search as you type capability, and an adjustable scale for typo tolerance (e.g. 3-4 letters off) are implemented out of the box. Dynamic Synonym Suggestions ensure that typing “flip flops” will deliver the same great results as “slides” in a search for summer footwear. Dynamic Re-Ranking identifies trends in queries and clicks and boosts them in results. Federated search and geo search capabilities ensure information is received from multiple sources at once and displayed on maps when useful. Algolia is constantly adding features to ensure searchers get a seamless and engaging experience no matter what they enter in the search bar.

Front end survival kit

Do you have, at minimum:

1. Instant, typo tolerant
   Search-as-you-type results

2. Results from all your
   Data sources, with smart, dynamic suggestions

For your developer team and business users

Algolia makes it easy for developers to deliver that frictionless UX with simple configurability and schema-less indexing. Onboarding resources, a pre-built dashboard, UI and prepackaged libraries make it quick and easy to build first-class front end components. No special search expertise? No problem.

With Elasticsearch, it’s another story. Engineering support is necessary to build new features. Typo tolerance, for example, requires extensive rule-setting and tedious busy work. 

It’s possible to work with a schema-less index in Elasticsearch, but it’s not the preferred option. Elasticsearch has built its libraries out slowly over time with supporting documentation that can be notoriously difficult to read. With the right developer skills, things can be done right, but updating inventory and correcting product specs is a headache and takes longer.

Back end survival kit

Do you have, at minimum:

1. Easy-to-use UX
   For your business users

2. Prepackaged libraries with simple configuration for your dev teams
   

Speed

Speed is non-negotiable in search. It’s the number one factor determining site success. Even a decade ago, a study revealed that just 100 ms of latency costs Amazon 1% of sales. Forrester Research found that fast search improves the user experience across channels, in-store, in hybrid retail, and on mobile. In a 2020 Deloitte study, a 100 ms speed increase on retail sites lowered bounce rates and increased mobile conversions by 8.4%.   

Algolia was built for speed. Every decision was made to eliminate latency at each layer of the stack. For example, instead of sorting results at query time, Algolia pre-sorts them to shorten response time. We use highly distributed architecture on bare metal servers, which speeds up processing at our end. So does splitting up search and indexing and assigning them different CPU priorities. 

Algolia Search is so fast. It’s instant.

Piotr Tuszewicki, Co-Founder @Noski Noski