Create a framework
Turned a broad business request into a decision framework.
Product Management Case Study
Hoozing Lead Management PRD
A lead management product specification that translated messy operational handoffs into standardized lifecycle logic, agent matching rules, scoring signals, and Customer Service execution cadence.
PRD System Snapshot
Lead lifecycle, matching, dashboard, scoring, and notification logic in one operating system.
The product work focused on turning manual judgment into clear rules: what the system decides, where users confirm, and which edge cases require guardrails.
8
Lead Status States
20+
Scoring Events
12
Pop-up UI Interactions
5
Core Modules
01 — Context & Ownership
Why this project strengthens my PM narrative
Point 1
At Hoozing, lead handling involved multiple teams, inconsistent follow-up, unclear ownership, and too much manual judgment.
Point 2
Approached it as a product problem: where decisions break down, what rules should be standardized, and how the system can guide better execution.
Point 3
Role went beyond documenting requests — framed the problem, mapped the lead lifecycle end-to-end, defined user interactions and permissions, and translated business complexity into clear logic for engineering and QA.
8
Lead Status States
20+
Scoring Events
02 — Product Decisions
The product decisions behind the PRD
The work was less about listing requested features and more about deciding how the product should behave under operational ambiguity.
Split automation from judgment
Clarified what should be automated, what required human confirmation, and what should remain manual due to business risk, timing, or operational edge cases.
Define operating logic
Defined the logic that drives the product: lead status transitions, follow-up cadence, ranking rules for agent matching, and operational guardrails.
03 — Core Features
Six core features turned lead handling into a managed product workflow.
Smart Lead Deduplication
Combines form submissions when same phone appears within 5-hour window. Duplicates flagged "Duplicate" and linked to canonical record.
Intelligent Agent Matching
Top-30 agents ranked by project→district fallback + weighted behavioral/performance score. Overloaded agents surfaced with tag.
Visit Calendar & Lifecycle
Per-lead calendar with CRUD, status tracking (Scheduled/Visited/Canceled), survey-on-completion flow. Cancellation/visited feed into agent scoring.
Kanban CS Dashboard
Daily Kanban for Customer Service sorted by lead priority, move-in urgency, pending follow-up. Card cadence by status (daily → every 5 days).
Push Notification & SMS
One-click dispatch to matched agents with auto-generated Vietnamese templates. SMS enforces 150-char limit with Zalo deep-link routing.
Weighted Agent Scoring
20-event scoring engine: Behavior (30%), Performance (60%), Listing Quality (10%). Balanced score ranks agents and manages overload thresholds.
04 — Business Logic
Status transitions and CS cadence made operational rules explicit.
Lead status states are lifecycle logic; dashboard cadence is sequential operational logic. Both stay tabular so teams can read them as rules, not marketing cards.
Lead Status States
| Status | Trigger |
|---|---|
| New | AutoAuto (default on creation) |
| Following Priority 1 | AutoAuto + Timeline (lead age ≤ 7 days) |
| Following Priority 2 / 3 | AutoAuto + Timeline (P2: 20–40 days, P3: >40 days) |
| Visiting | AutoAuto + Event + Manual (4-day window around scheduled visit) |
| Negotiating | EventEvent (agent trigger) |
| Customer No Reply | EventEvent (3× checkbox by CS, confirmation dialog required) |
| Customer Drop | ManualManual (mandatory drop reason + drop date) |
| Cannot Serve | ManualEvent (≥3 agents change) + Manual |
| Closed Deal | ManualManual (creates Deal record in background) |
New
Auto
Auto (default on creation)
Following Priority 1
Auto
Auto + Timeline (lead age ≤ 7 days)
Following Priority 2 / 3
Auto
Auto + Timeline (P2: 20–40 days, P3: >40 days)
Visiting
Auto
Auto + Event + Manual (4-day window around scheduled visit)
Negotiating
Event
Event (agent trigger)
Customer No Reply
Event
Event (3× checkbox by CS, confirmation dialog required)
Customer Drop
Manual
Manual (mandatory drop reason + drop date)
Cannot Serve
Manual
Event (≥3 agents change) + Manual
Closed Deal
Manual
Manual (creates Deal record in background)
CS Dashboard Follow-up Cadence
| Lead Status | Card Frequency |
|---|---|
| New | Every day |
| Following Priority 1 | Every day |
| Following Priority 2 | Every 3rd day |
| Following Priority 3 | Every 5th day |
| Visiting | Every day |
| Negotiating | Every 2nd (Rent) / 5th (Buy) |
| Closed / Dropped | Removed from board |
New
Every day
Following Priority 1
Every day
Following Priority 2
Every 3rd day
Following Priority 3
Every 5th day
Visiting
Every day
Negotiating
Every 2nd (Rent) / 5th (Buy)
Closed / Dropped
Removed from board
05 — Scoring Engine
A weighted scoring engine balanced behavior, performance, and listing quality.
Behavior
30%- Poor communication rating (<3★): −1
- 3× no-reply in one lead: −4
- Poor visit quality (<3★): −2
- Property mismatch on photos: −2
- Post-deal poor support: −4.5 max
Performance
60%- Visit cancel (personal): −1 per event
- Mismatched properties sent: −0.25
- Visit rate below 50%: −0.5/1%
- Negotiation rate below 75%: −0.25/1%
- 5+ unaccepted target popups: −2 each
Listing Quality
10%- 50%+ reminders ignored: −0.05/1%
- No listing updates in 10 days: −0.25
Formula: Final balanced score = Σ (Metric Score × Weighting), normalized against worst/best case across all agents.
06 — Scope & Scale
The PRD covered enough operational surface area to behave like a system.
8
Lead Status States
End-to-end lifecycle coverage
20+
Agent Scoring Events
Weighted performance inputs
12
Pop-up UI Interactions
Confirmation and action flows
5
Core Modules
Leads, Visits, Dashboard, Scoring, Notifications
07 — Skills Demonstrated
The case study shows product judgment across workflow, algorithm, edge case, and role design.
Complex Workflow Decomposition
Broke down intricate state machine (lead ↔ agent ↔ visit status) into clear, testable rules distinguishing automatic, event-driven, and manual transitions.
Algorithm Design for Ranking
Multi-criteria agent scoring with time-decay windows, balanced weighting, and fallback logic for district/project matching.
Edge Case Thinking
Explicitly documented all "does not allow" cases, such as agent assignment blocked when lead is Closed/Dropped.
Cross-functional Specification
PRD covers UI interactions, backend business rules, cron job logic, database migration fields, and SMS character limits.
Localization & Market Context
Vietnamese notification/SMS templates with dynamic field injection. Understood local real estate conventions (VND/USD pricing, 6-month minimum rental policy).
User Role & Permission Design
Role-specific views (Customer Service vs CS Leader vs Admin) with conditional display logic and action permissions.
Back to Portfolio
Explore the rest of the work.
Return to the homepage for the full portfolio, PM journey, and supporting work.
Back to Portfolio