Personalization is showing the same items but in different order for different users.

The order of posts in FB, photos in Instagram, products in Amazon, and search results in Google is personalized for each visitor, as it directly affects user engagement: click rate and conversion. We've done 50+ a/b tests in different ecommerce verticals to confirm it.

If you have items that are presented to a user in a specific order, you can personalize this order to improve your product's KPIs.

Metarank can run on-premises or in the cloud, making it easy to get state-of-the-art ML technologies by teams having little to no ML experience.

Building an in-house personalization solution can take around 6 months by an experienced team, and only large companies can afford it.

With Metarank, you can get up and running with personalization in days instead of months and be in control of data privacy, optimization goals and how Metarank is integrated into the infrastructure with minimal vendor-lock.

We have a built a Demo which showcases how you can use Metarank in the wild.

The Demo utilizes Ranklens dataset that we have built using Toloka service to gather user interactions.

Application code can be found here and you can see how easy it is to query Metarank installation to get real-time personalization.

Check out the step-by-step Tutorial of running Metarank Demo locally.

• Overview of the way it can be integrated in your existing tech stack.
• Configuration
• Run & Deploy
• Tutorials
• Contribution guide
• License

We recommend running Metarank with Docker.

If you would like to try it natively, here are the requirements:

• Linux or MacOS on x86. Apple M1 and Windows support are coming soon.
• JVM 11

This project is released under the Apache 2.0 license, as specified in the LICENSE file.