Revolutionizing Rentals with GenAI-powered Search for 35% Increased Conversions
About the Client
The client is a leading online marketplace that disrupted the pool and sports venue rental industry. They empower homeowners to list their private amenities (pools, courts, gyms, etc.) or entire homes for hourly rentals to individuals or groups. Reservations, communication, amenity selection, and payment are all handled seamlessly through their mobile app and website, currently operating in the US, Canada, and Australia.
Their Challenges
In today's digital marketplace, providing an intuitive and efficient user experience is crucial for success. The client, a leading online platform for renting pools and sports venues, identified several challenges hindering their ability to deliver optimal search functionality for their users:
Limitations of Keyword-Based Search
The existing keyword-based search system struggled to capture the nuances of user queries, particularly those involving visual aspects like "sunset view" or desired amenities like "spacious seating areas." This resulted in users potentially missing out on ideal rental properties that matched their preferences but lacked specific keywords in the descriptions.
Inaccurate Search
Results
The keyword-based approach often led to irrelevant search results, frustrating users and hindering their ability to find their perfect pool or court rental. This inefficiency could discourage users from continuing their search and potentially result in missed booking opportunities for The client's rental partners.
Limited Ability to Discover Unique Offerings
The traditional search functionality restricted users' ability to discover unique amenities or hidden gem locations that may not have been explicitly mentioned using keywords. This limited the overall user experience and potentially prevented them from exploring the full range of rental options available on The client's platform.
The Solution
The client partnered with Searce to develop an AI-powered search engine. Going beyond traditional text-based methods by incorporating both textual data and image data in the search process, our solution offers a more comprehensive user experience.
Here's a breakdown of the implemented solution:
Data Collection and Representation
The client provided property data, including descriptions, features, and corresponding images for their listings. Searce then leveraged Vertex Embeddings API to convert this raw data into a "vectorized representation." This embedding process combines text and image data information, creating a richer search representation that captures the essence of each property.
IMPACT
Enhanced Relevance
Richer data matching leads to more relevant search results beyond keywords. with users discovering hidden gems and 90% increased user satisfaction
IMPACT
Increased Bookings
Perfect property matches translate to a 35% increase in booking rates on the client's platform
Efficient Search with Vertex Vector Search
The system utilizes Vertex Vector Search to perform efficient searches. Users can filter their searches by city and property type, and the system retrieves the most relevant properties based on these filters and the user's query. This search process considers the properties' textual content and visual aspects, ensuring highly relevant results for users.
Scalable Deployment with Cloud Run
Cloud Run, a serverless platform on Google Cloud, was chosen for deployment. This ensures scalability to handle increasing search demands while maintaining ease of use. The client can access the search functionality through a user-friendly REST API, enabling seamless integration with their existing systems.
IMPACT
Scalable & Future-Proof
Built on Google Cloud, the solution scales with growth. Multi-modal search lays the groundwork for future innovation
Beyond Search Functionality: Searce acted as a one-stop shop for the client's account and billing, including providing a sub-account tailored specifically to their requirements. This simplified the billing process and streamlined their admin experience.
Success Story Highlights
This leading booking website's AI-powered search engine using Vertex AI cut search response times in half (from 300ms to 150ms) for a smoother user experience. Advanced filtering combined with scalable Cloud Run delivers highly relevant results, ranking properties based on user preferences. This translated to a 250% increase in booking conversions, from 10% to 35%. This also increased user satisfaction, by 90%.
more case studies
