← All Articles
Project StoryProject StoryReal Estate

Build log: the map-first portal that doubled Vertex Realty's enquiries

Kunal Desai
Kunal Desai
Engineering Lead · 1 Mar 2026 · 9 min read
Build log: the map-first portal that doubled Vertex Realty's enquiries
This is the build log for
Vertex Realty - 2x enquiries in 90 days
View case study →

Vertex Realty came to us with 4,000+ listings and a portal that treated its map as an afterthought. Here's how we rebuilt it.

WEEK 1-2
Discovery: shadowing brokers

We spent the first two weeks in their offices, listening to enquiry calls and following brokers on site visits. The pattern was unmistakable: buyers asked the same five questions on every call, and brokers spent most of their day answering questions a listing page could answer. That gave us our thesis - win the search results, then remove every step between "found it" and "visited it."

WEEK 3-5
Design: the map is the homepage

Every portal we benchmarked treated the map as an afterthought bolted onto a list. We flipped it: the map IS the browse experience, with faceted filters riding on top. Prototype tests with eight buyers confirmed it - people navigate neighbourhoods, not sort orders.

WEEK 6-12
Build: the clustering problem
The Challenge

With 4,000+ listings, naive client-side map markers froze mid-range phones for seconds on every pan. Clustering libraries helped but re-clustered visibly on every zoom, which felt broken.

What Worked

We moved clustering server-side, keyed by zoom level and viewport bounds, and cached cluster tiles at the edge. Panning and zooming both stayed under 16ms.

The best schema decision we made: treating every listing as a content page first and a database row second. SEO paid the bills here.

Three months after launch, enquiries had doubled and organic search had become the single largest acquisition channel.

#nextjs#maps#seo#real-estate
Kunal Desai
Written by Kunal Desai
Engineering Lead at MaxielaTech. Engineering Lead at MaxielaTech. Ships web platforms and still thinks in SQL.
Building something similar?
Tell us about it - free analysis, honest timelines.
Start a Project