First decision: where do listings come from?
Everything in a real-estate build flows from the listing-source question. The options, honestly ranked: manual entry (you or the agency enters each property, fine for a single agency with dozens of listings, dead on arrival for a portal ambition); MLS/IDX feeds (the US/Canada brokerage data system, licensed access through IDX providers, monthly fees, display rules that constrain design); XML/CRM feeds (agency software exports, the common case outside North America, where the build ingests feeds from source systems on a schedule); scraping portals (against their terms, declined here, said plainly). The source decision sets the data model, the refresh cadence, the budget, and the legal surface, it is the first scoping conversation, before design.
Search: how buyers actually filter
Property search is faceted search with strong opinions: location (with map behavior, draw-a-shape searches are the feature users remember), price bands, property type, beds/baths, and the vertical-specific filters (plot size, floor number, Furnished, gated-community-only). The map experience is the emotional core, listings without a working, fast map feel broken to property buyers, and it is also the heaviest component (the Maps integration page covers the speed and billing discipline). Saves and alerts (email me new matches for this search) are the retention mechanism that separates a lead tool from a brochure.
Lead routing: the reason the site exists
A property site is a lead machine, and the half that determines ROI is invisible: what happens after "Request a visit". Routing rules (which agent, by property/region/language), speed (leads decay in hours, auto-acknowledge instantly, human follows fast), and the feedback loop (lead status back into the system, so the site knows which listings convert) turn traffic into revenue. CRM integration rides the standard patterns, the CRM integration page covers the plumbing; WhatsApp routing matters enormously in some markets (the WhatsApp Business page covers that rail). Quotes that stop at "contact form" are quoting half a product.
SEO: the compounding asset (and its traps)
Real estate search traffic lives in long-tail structure: "3BHK flats in [locality]", "villas for sale near [landmark]", thousands of indexable surfaces from listing data + locality pages done right. Done wrong, it is doorway-spam that Google demolishes (thin auto-generated locality pages are the classic casualty, the anti-doorway rule applies twice over here). The honest architecture: listing pages with real schema markup, locality guides with genuine content, faceted URLs that stay crawlable without exploding into duplicate variants. Feeds-based sites add a duplication discipline problem (the same property on forty agency sites), canonical strategy decided deliberately, not by plugin default.
What the build contains
- Listing ingestion, per the source decision: manual admin, IDX provider wiring, or feed importers with dedup and refresh schedules
- Search + map experience, facets, saved searches, alerts; fast on real data volumes
- Listing pages, galleries, particulars, schema markup, agent contact with routing
- Lead machinery, routing rules, auto-acknowledgement, CRM handoff, status loop-backs
- Agent/agency layer, profiles, their listings, their leads (for multi-agency builds)
- SEO architecture, the locality/listing structure, schema, and the canonical discipline feeds demand
The service framing for agency-class builds sits at the real-estate industry page; bands ride the published context and the calculator prices a configuration. Bring the listing source answer and the lead flow, the brief turns it into a scoped quote.