How Exelerate helped Rentd launch a multi-platform rental platfrom in London & Dubai within a budget
Currently operating in London and Dubai, both booming markets in the rental space industry, Rentd’s goal is to revolutionize the long-term letting market. Outside of Rentd the process of discovering, viewing, negotiating, getting approved and closing on a property is tedious at best. The whole process requires a human interaction with multiple people involved and takes weeks to finalize. Rentd extrapolates all of that complexity and makes it possible for both tenants and landlords to close on a deal completely through the platform within days, as opposed to weeks.
These are not just a group of guys that can pump out code, they really know business and provide you with the expertise you need to take your startup to the next level, as a non-technical founder this was key.
Rentd connects tenants who wish to rent an apartment with landlords who rent out apartments. The platform currently has coverage right across London and Dubai with full-spectrum specialisms ranging from listing a property to signing a contract and moving in. Its offer is particularly attractive to customers who want to move quickly and avoid the manual and physical process.
“Renting a property has never been so easy. Our app offers an end-to-end journey for renters and landlords. Everything from finding a property to signing a contract can be completed in one day”, says Ahmed, the founder and managing director at Rentd.
Rentd approached us in 2020 with only an iOS app built with “Swift” and an Admin Panel built with an old version of Angular. The requirements were to expand the product by building a Web App and an Android App, as well as customizing the platform for the Dubai market. Apart from that, the whole journey needed to be automated, including the signing of the contract and the payment process of the first rent, which was not included in the iOS app.
Another big challenge was building a powerful search, allowing the users to easily and quickly find properties, since the existing search was slow and inaccurate due to lack of proper architecture.
Building a software product across all platforms (Web, iOS, Android) for different markets requires a good plan in terms of time-to-market, velocity, scalability and robustness.
We started with a Solution Architect who was creating the project plan (software architecture, cloud infrastructure, release plan and team structure) based on the resources we had and the company’s OKRs.
We needed to find an acceptable ratio between the time to market and the robustness and scalability of the product.
To do so, we divided the plan into two main parts:
1. Technology and tools
The new milestone included an Android app, so there was no dilemma about choosing React Native as a technology for building cross-platform applications, which allowed us to reuse the business logic of the web app as well as get the iOS app built with a very low effort.The Admin Panel was the only app that was left outdated and built with different technology, an old version of Angular, that made us think and decide whether to keep it and iterate on the existing app or build a new one using NextJS. We went with building it from scratch and thus, we ended up having 4 apps using React and Typescript as a main technology stack that gave us the flexibility to easily scale and maintain the product.
The business logic was wrapped in an HTTP Cloud Function which was totally wrong when it comes to concurrency, cold-starts and costs. We improved the code, containerized and deployed it using Cloud Run as a service for running Serverless containers. The flexibility for setting a minimum and maximum number of instances helped us avoid cold starts and scale automatically. Additionally, by increasing the number of concurrent requests, response times were reduced.
We kept the background functions, but changed the location to eu-west (previously us-east) since the business has been operating in London, as well as improved the quality of the code by using the best practices and rules we’ve built in the past 5 years running a Product Studio. Github.
Firestore (NoSQL) was selected as a primary database, which was not adequate at all for buidling the search, since it contains large criteria and needed to be an advanced search with a lot of filters to give the user the possibility to quickly find properties with the desired parameters. We chose Elasticsearch as a most powerful free and open search engine available.
The images of the listings are in high resolution, which impacts the load time. Instead of resizing and duplicating each image in the storage, we built an image service using Python that resizes images on the fly based on the user’s screen.
Rentd became a fully automated rental platform operating in London and Dubai. It offers its services across all platforms (Web, iOS, Android).The software is maintainable, scalable and robust. It uses the latest technologies available on the market.
Exelerate helped Rentd to be able to easily expand to new markets with very low effort and zero cost for development.