How We Built a LinkedIn for Real Estate

The Coordination Problem

Real estate transactions require coordination among specialists. Agents find buyers and sellers. Lenders provide financing. Inspectors evaluate properties. Title companies handle closing. Photographers create marketing materials. Each professional knows others in their specialty but struggles to find collaborators in adjacent fields.

We built a platform connecting these professionals—a LinkedIn for real estate. The technical challenges were fascinating: modeling complex professional relationships, building discovery algorithms, and creating verification systems that balance accessibility with trust.

The Taxonomy Problem

Real estate has dozens of specializations, and the boundaries are fuzzy. Is a "real estate investor" the same category as a "real estate syndicator"? Is a "transaction coordinator" a subset of "real estate agent" or a separate profession?

We settled on a two-level taxonomy: broad categories (transaction, finance, legal, inspection, marketing) containing specific professions. Each profession declares which other professions it commonly works with. An agent typically needs lenders, inspectors, and photographers. A lender needs agents, appraisers, and title companies.

This relationship graph powers discovery. When an agent searches, lenders rank higher than architects—not because lenders are better, but because the relationship is more common.

Geographic Scope

Professionals operate at different geographic scales. Real estate agents are hyper-local—they know one city or one neighborhood. Mortgage lenders cover regions or states. Investors might operate nationally.

The discovery algorithm must respect these scopes. Showing a New York agent to a California buyer wastes everyone's time. Showing a national investor to either might be relevant.

We model service area as a union of geographic primitives: radius from a point, set of counties, set of states, or nationwide. Discovery filters by overlap—a query from San Francisco returns professionals whose service areas include San Francisco.

Trust Layers

Professional networking requires trust, and trust has levels.

Email verification confirms the user is real. Identity verification confirms they are who they claim. Professional verification confirms their licenses and credentials. Premium verification adds background checks and references.

Each level unlocks features. Unverified users can browse. Verified users can connect. Professionally verified users appear in search results with badges. Premium users get highlighted placement.

The verification process must balance friction against trust. Too easy, and the platform fills with spam. Too hard, and legitimate professionals don't bother. We found the sweet spot varies by profession—licensed professions need license verification, while unlicensed roles (photographers, stagers) use portfolio and reference verification instead.

License Verification

For licensed professions, we verify against state licensing databases. Many states provide APIs. Others require document submission and manual review.

The API path is fast and authoritative. Submit license number and state, receive confirmation or denial. The manual path is slow but handles edge cases—expired licenses pending renewal, licenses from states without APIs, international credentials.

Verification isn't binary. A verified license that expires becomes "verification expired." A license that can't be confirmed becomes "pending review." The UI communicates these states clearly so users understand what verification means.

The Feed Algorithm

Users see a curated feed of potential collaborators. The algorithm combines multiple signals.

Profession relevance matters most. If you're an agent, seeing lenders and inspectors is more useful than seeing other agents. The profession relationship graph weights this heavily.

Geographic proximity matters for local professions. An agent in Dallas wants to see Dallas-area lenders, not Houston ones.

Network effects matter. Professionals connected to your existing connections are more likely to be trustworthy and relevant. Second-degree connections get a relevance boost.

Activity matters. Professionals who respond quickly and engage regularly are more valuable connections. Inactive accounts sink in rankings.

These signals combine into a score that determines feed ordering. The weights were tuned through user feedback and engagement metrics.

Connection Dynamics

Connections can be direct or introduced. Direct connections happen when one professional reaches out to another. Introduced connections happen when a mutual connection facilitates the introduction.

Introductions carry more weight. When someone you trust vouches for someone new, that trust transfers partially. The platform makes introductions easy—one click to offer connecting two of your contacts.

Connection requests include context: "I'm looking for an inspector for a closing next week" communicates differently than "I'm building my professional network." Context helps recipients prioritize and respond appropriately.

Real-Time Communication

Once connected, professionals message through the platform. Messages need to be instant—real estate moves fast, and delayed communication loses deals.

Real-time messaging requires persistent connections, typically WebSockets. The challenge is scaling—each connected user holds a socket, and real estate professionals are often online during business hours simultaneously.

The solution is presence management with graceful degradation. Active sessions get real-time updates. Inactive sessions get push notifications. Offline users get email summaries. The system adapts to user state without requiring user configuration.

Discovery Versus Serendipity

Good discovery means finding the professional you need. Great discovery means finding professionals you didn't know you needed.

The feed intentionally includes some low-relevance results. A random photographer appearing in an agent's feed might spark an unexpected collaboration. Pure optimization would filter these out; strategic randomness preserves serendipity.

This is a deliberate design choice. The platform isn't just a search engine—it's a professional community. Communities benefit from unexpected connections.

The Network Effect

Platform value grows with network size, but not linearly. The first hundred users provide little value to each other—the network is too sparse. The first thousand provide substantial value—most queries find relevant results. Beyond that, incremental users add incremental value but also incremental noise.

Managing this growth curve means prioritizing quality over quantity early. Verification requirements filter low-quality signups. Invitation systems create social accountability. Geographic focus concentrates users rather than spreading thin.

The goal is network density: enough professionals in each category, in each location, to make discovery reliable. Sparse networks disappoint users; dense networks create value.

The Broader Insight

Professional networking platforms solve coordination problems. The technology is secondary to the design—understanding which relationships matter, how trust forms, and what makes discovery useful.

The code handles taxonomies, verification, feeds, and messaging. But the real work is understanding real estate professionals: their workflows, their relationships, their trust patterns. Technology implements that understanding; it doesn't replace it.