Vireo Properties

Problem

Belize property discovery is fragmented across Facebook, WhatsApp, spreadsheets, and agent memory. Buyers struggle to compare trustworthy information, while agents manage listings, inquiries, follow-ups, and parcel details across disconnected tools. The central product problem is trust: the software must show useful facts without implying title verification that has not happened.

My role and ownership

I founded the product and built it end to end: domain discovery, architecture, data modeling, interface design, GIS workflows, implementation, deployment, and the trust boundaries visible to buyers and agents.

Constraints

• Real listing data arrives with uneven structure and provenance.
• Belize property is advertised in both BZD and USD.
• Surveyed subdivisions share approvals, roads, and source documents across many lots.
• The product must distinguish agent representation from independent title verification.
• A small operating team needs one maintainable system, not a collection of bespoke tools.

Architecture or walkthrough

The public Next.js surface reads typed listing data from Postgres, while the operator surface connects listing management, leads, agents, and sales intelligence. A separate Python GIS workflow turns survey material into queryable parcel geometry. Consent-gated measurement and WhatsApp handoffs sit outside the core data boundary.

The walkthroughs above show the public product, the lot-level subdivision model, and the buyer qualification flow.

Important decisions

1. Subdivisions are parent entities

Context: One approved subdivision can contain dozens of lots that share a survey, approvals, and road geometry.

Choice: Model a subdivision once, with lot children carrying their own dimensions and availability.

Reason: Duplicating shared provenance across independent listings creates many places for the trust-critical source data to drift.

Tradeoff: Search, cards, maps, and the CRM all need to understand two related entity shapes.

2. Currency of record stays attached to the listing

Context: Belize pegs BZD to USD, but a seller's advertised number is still a fact about the source listing.

Choice: Store the advertised amount and currency, then derive the secondary display value.

Reason: A rounded conversion should not silently replace the claim a buyer may later bring into a negotiation.

Tradeoff: Every price surface needs currency context, and any future change to the peg requires a real exchange-rate layer.

3. Trust boundaries appear beside the claim

Context: The platform presents prices and property details but does not independently verify title.

Choice: Put representation and verification limits beside the relevant property information, with a clear recommendation to retain an independent Belize attorney.

Reason: A footer disclaimer is technically present but practically invisible at the moment a buyer forms an impression.

Tradeoff: Explicit caveats can reduce short-term conversion.

4. Measurement is consent-gated

Context: Advertising attribution can share visitor and lead signals with external providers.

Choice: Keep measurement off until explicit consent, with decline presented as an equal action.

Reason: Trust and compliance matter more than complete attribution for a product supporting high-value decisions.

Tradeoff: Marketing data remains incomplete.

Defensible outcome

Vireo is live with real inventory and an active agent relationship. It connects public discovery, buyer qualification, listing operations, and parcel workflows while making provenance limits explicit. It is still pre-revenue, and no transaction is attributed to the platform here.

What I would improve

I would establish a normalized location model for districts, towns, and subdivisions from day one. Early hardcoded routes accelerated the first release, but migrating location relationships after real inventory exists is riskier than making them first-class at the start.