Case study · Real estate
Realty Canada, a national real estate portal
A nationwide buy / sell / rent property portal with agent profiles, advanced search, saved-search alerts and a brokerage-grade listings backend, designed to compete with the incumbents on speed, search quality and a less cluttered experience.
The challenge
What we had to solve.
Listings data spanned multiple regional MLS feeds, each with different fields, refresh cadences and licensing terms.
Buyers and sellers expected a modern Zillow-style experience, fast search, map exploration, saved searches, price-drop alerts, while agents needed a no-friction way to manage their own listings and leads.
The product had to scale from a single launch market to national coverage without rebuilding the architecture twice.
Our approach
How we tackled it.
We started with a 3-week discovery to map the data sources, regulatory constraints (CREA/REALTOR rules, provincial nuances) and the agent / buyer / seller journeys.
We built a normalisation layer that pulls from each regional feed, harmonises the schema and writes into a single Postgres index, so the front-end works the same nationwide.
Search was rebuilt on a typo-tolerant engine with map clustering, faceted filters (price, beds, sq ft, lot size, year built, days on market) and a saved-search service that fires price-drop and new-listing alerts by email and push.
The agent dashboard ships listings, lead inbox, saved-search insights for their region and a profile editor. Brokerages get team accounts with role permissions.
Deliverables
What we built.
Specific, named outputs, not vague "strategy".
Public portalListing pages, map search, saved searches, alerts, mortgage estimator and a clean mobile-first UI.
Agent dashboardListings CRUD, lead inbox, performance metrics and a calendar-linked tour-request widget.
MLS normalisation pipelineRegion-aware ingestors, deduplication, geocoding fallback and an audit log of every feed change.
Notification engineDaily-digest and instant-alert workflows for price drops, status changes and new matches.
Marketing siteAn SEO-tuned content layer, neighbourhood guides, market reports and agent profiles, that brings buyers to the portal.
Outcomes
What it returned.
- Country-wide listings coverage by month 9, with three regional launch waves rather than a single big-bang go-live.
- Search-result-to-listing-view rate measurably higher than the incumbent benchmark on identical queries, thanks to typo tolerance and faceted filters.
- Agents who adopted the dashboard self-served 80%+ of their listing updates, fewer support tickets, faster listings on market.
The takeaway
What we learned.
A real estate portal lives or dies on its data layer. Spend the budget on a clean normalisation pipeline before you touch the UI, every shortcut you take there comes back as a customer-support ticket six months later.
Services used
How we got it done.
Related work