Homegate’s fast and modern search experience helps users find their dream home

Homegate, founded in 2001, is a well known Swiss real estate marketplace and a brand of SMG Swiss Marketplace Group. SMG is a pioneering network of online marketplaces with a portfolio including brands across several business units in real estate, automotive, and general marketplaces, as well as financing and insurance. 

Homegate’s team has greatly modernized its tech stack over the past several years — it was completely rewritten from a monolith application several years ago, and it now consists of microservices, deployed in the cloud on AWS. The team heavily utilizes serverless to avoid dealing with servers and other physical infrastructure. Homegate’s front end is written mostly in Vue.js. Aside from AWS, Homegate also uses other cloud services for more specialized tasks. Elasticsearch is one example — helping the team create a modern search experience for their visitors every day.

When a Homegate user cannot find the property that fits their search criteria, instead of regularly checking the platform for the new listings, they can enter their search criteria and email address to receive a notification as soon as a matching property is listed on the platform. Users can set customized alerts across search criteria like property type, location, price, and more. This is one of the most popular services Homegate provides, and it is available on the website and in apps — similar features are also offered across other SMG platforms.

As a result, Homegate manages a high volume of search alerts, with even more notifications each month, requiring near real-time notifications to users via emails and mobile app pushes. To do this, the Homegate team needs a lightning fast and highly scalable tool to store indexed search queries (i.e., alerts) from the users — and that is why Elasticsearch is at the heart of the engine.

When a user creates a search alert, Homegate:

1. Persists it in its AWS infrastructure

2. Indexes it in the search alert index

3. Leverages Elasticsearch to store the indexed search queries

Then when an advertiser posts a property on the platform:

1. The search alert service receives the notice about the property

2. It creates a percolate query for the property (Percolate queries check all saved search criteria against this new property and return those that the property satisfies.)

3. And finally, it executes search in the Elasticsearch, finding all the matching user queries and notifying the users

Elasticsearch serves as the perfect solution for Homegate’s modern architecture and the services it offers. One particular feature in Elasticsearch that the team found helpful is index lifecycle management. It is a good practice to have continuous deployment, especially with microservices architecture. It is also crucial to have good test coverage and an easy setup for development, so engineers can try out their changes. With index lifecycle management, Homegate’s Engineering team sets the playground indices to be deleted automatically after several days of inactivity. The feature simplifies management of the playground and tests deployments, removing the need to manage these deployments explicitly.

The release and timing of any features or functionality described in this post remain at Elastic's sole discretion. Any features or functionality not currently available may not be delivered on time or at all.