Dada Room

Scalable Marketplace Capabilities (Discovery · Contact · Trust)

Outcome

Built a role-aware marketplace foundation—unified discovery across inventories, quality-gated contact mechanics, and systemized patterns—so users could progress from exploration to qualified contact with confidence.

A marketplace platform designed to help users find the right place and the right fit—across evolving roles and sub-markets.

Role

Co-founder & Head of Product & Design (platform strategy, UX systems, growth experimentation, monetization surfaces)

Year

2012–2018 (time-series results shown 2015–2018; product KPIs highlighted 2018)

Industry

Marketplace · Housing / Rentals

Team / Stakeholders

Startup team (product/design + engineering), plus cross-functional partners across operations, marketing, and customer support.

TL;DR

Problem

As Dada Room grew, the marketplace evolved from a simple two-sided model into multiple sub-markets with users switching roles over time. That complexity increased decision friction, lowered trust, and made “contact” noisy—putting marketplace health at risk.

Strategy

Build scalable platform capabilities—<b>unified discovery across two inventories</b> (people and places), standardized decision units (cards + detail templates), and quality-gated contact mechanics that protect trust while enabling monetization. Systemize patterns to increase iteration speed as segments and journeys multiplied.

Impact

Platform scale and monetization grew over multi-year periods (users/listings, premium subs, premium revenue). In 2018, “qualified contact” proxies improved materially (first-contact message volume and contact starts per active user), alongside higher listing creation conversion (supply activation). Measurement method: time-series dashboards and cohort cuts where available; early signals explicitly labeled when attribution is incomplete.

Metric Tree & Success Criteria

North Star (NSM)

Proxy tracked as “Primer contacto” message type (first-contact messages).

Activation

Listing creation conversion (supply activation). Proxies: % active users creating a listing; % sign-ups creating a listing.

Engagement

Messages per active user;

listing exploration depth (search → detail usage).

Retention

Repeat qualified contact starts (7/30d);

return to search/revisit listings (tracked as framework where full data isn’t available).

Validation

Time-series trends + cohort/segment cuts (inventory type, role/intent) using internal dashboards sourced from analytics and DB extracts; “early signals” where full attribution isn’t available.

Metric tree defining the North Star (qualified contact starts) with activation, engagement, retention, and trust/quality guardrails. — Success criteria centered on qualified contact—validated through time-series outcomes, cohort cuts, and early signals when attribution is incomplete.

Research insights

A single “two-sided marketplace” mental model broke over time. The platform expanded into sub-markets: seekers, hosts (spare room vs entire place), short-term use cases, compatibility-led seekers, and multi-listing managers. Users could shift between these roles, sometimes repeatedly.

Discovery had to support two inventories without resetting intent. Users don’t search for “a thing”—they search for a match. Switching between people (roomies) and places (properties) needed to preserve location, budget, and preferences.

Contact is the real conversion point, but it must be protected. In a housing marketplace, low-friction contact can quickly become spam. “Qualified contact” required guardrails, clear expectations, and trust cues across the journey.

Multi-session behavior is the default. Housing decisions unfold across multiple sessions, so continuity (revisits, conversation pickup, saved exploration) matters as much as first-session conversion.

Diagram mapping marketplace segments and role transitions across seekers, hosts, and multi-listing managers. — As the marketplace expanded into sub-markets and role switching, role-aware journeys and tailored packaging became a platform requirement.

Design Decisions

Why

Unified discovery across inventories via a filter switch (Roomies ↔ Properties). Why: preserve intent and reduce cognitive load while pivoting inventory type. Trade-off: higher taxonomy complexity and tighter component governance. Expected impact: more confident progression to qualified contact starts.

Trade-offs

Standardize decision units (cards + detail templates) across both types. Why: improve comparability and reduce ambiguity at scale. Trade-off: limits one-off layouts; requires content rules. Expected impact: higher search→detail engagement and cleaner intent actions.

Expected impact

Treat contact as a quality-gated capability (not a feature). Why: protect trust and reduce noise while enabling monetization surfaces. Trade-off: added friction; optimize for quality, not raw volume. Expected impact: better NSM quality and stronger two-way conversation proxies.

Platform capabilities map covering unified discovery, contact mechanics, trust governance, lifecycle loops, and systemization for speed. — Scalable capabilities kept discovery and contact consistent across inventories while enabling quality guardrails and faster iteration.

Solution

Patterns / Flows / Modules / Templates

Unified search + filters that allow switching inventories without restarting the journey.

Responsive / Mobile-first / A11y

Consistent listing cards + detail templates that keep “what it is / why it fits / how to contact” predictable.

Tie-back to business objectives

Role-aware contact mechanics (Premium vs Non-Premium paths) guiding users from discovery to qualified contact starts.

Hi-fi user flow showing unified discovery and branched contact mechanics for Premium versus Non-Premium users (contextual messaging vs paywall). — A unified search system leads into role-aware contact mechanics—Premium users message contextually, while Non-Premium users see an upgrade gate—balancing qualified contact with quality and monetization.

Results

Metric 1

Scale (2015→2018):

Metric 2

Qualified contact (2018 proxy):

Metric 3

Activation (2018):

Method 1

Registered users 53,566 → 760,863 (+1320%); total listing base 36,220 → 558,823 (+1443%).

Method 2

Contact starts per active user 0.37 → 0.90 (peak), using “Primer contacto” as the proxy (time-series / early signals).

Method 3

% active users creating a listing 20.1% → 44.6% (peak) (time-series / early signals).

Results scorecard summarizing time-series outcomes and early signals for qualified contact, supply activation, and monetization. — Time-series outcomes show platform scale, rising qualified contact proxies, improved listing creation conversion, and supporting monetization signals—reported without feature-level attribution.

Systemization & Reuse

Systemized reusable UI patterns (cards, detail modules, contact CTAs) to reduce rework and speed iteration as roles, inventories, and journeys expanded.

UI kit and reusable patterns used to standardize cards, listing detail modules, and contact CTAs across inventories. — Systemization created leverage—reusable patterns reduced rework and enabled faster iteration as roles, inventory types, and sub-markets expanded.

What I'd Do Next

Add a listing/profile quality score to improve ranking and reduce noise;

Strengthen measurement around two-way conversations as a quality KPI downstream of contact starts; Expand role-aware defaults (seekers/hosts/managers) to improve discovery efficiency.

Credits

Product strategy, platform UX, and design direction: Juan Avalos · Team: engineering, ops, marketing, support.

Tools

Tools varied over time (Sketch/Figma + analytics dashboards/internal reporting).