Product Portfolio — How the Three Products Interlock¶

The three products¶

Product	Audience	Status	Revenue model
PropPie Fractional	Retail investors (₹25K-₹5L tickets)	Live	Transaction fees + management fees
PropPie Broker	Retail / HNI / NRI property buyers	Design	Open question — subscription / freemium / per-report
PropPie Analytix	Developers, brokers, institutions	Design	Open question — tiered SaaS subscription

The shared data foundation¶

All three products sit on one canonical data layer. This is the architectural thesis:

flowchart TD
    subgraph Sources[Raw Data Sources]
        RERA[MahaRERA<br/>40K projects]
        IGR[IGR Index-II<br/>transactions]
        GR[Govt Resolutions]
        BHU[MahaBhulekh]
        GIS[GIS Layers]
        NEWS[News + Social]
        INT[Internal - Fractional DB]
    end

    subgraph Pipeline[Pipeline - Vishal]
        EXT[Extraction + OCR + LLM]
        ER[Entity Resolution]
        QP[Quality Passport]
    end

    subgraph Store[Canonical Attribute Store]
        RAW[~90 raw attributes]
        DER[~50 derived attributes]
        LIN[Lineage + Confidence]
    end

    subgraph AI[AI Layer]
        RAG[RAG over documents]
        SCORE[Scoring engines]
        SIM[Simulation - Monte Carlo]
        NLP[NLP - GR classifier + sentiment]
        LLM[LLM - narratives + explanations]
    end

    subgraph Products[Products]
        FRAC[PropPie Fractional<br/>AI overlay on asset cards]
        BROKER[PropPie Broker<br/>Conversational B2C]
        ANALYTIX[PropPie Analytix<br/>B2B dashboards]
    end

    Sources --> Pipeline --> Store --> AI --> Products
    INT --> Store

Why this architecture (not three separate products)¶

Data is the moat. Building the MahaRERA + IGR + GR fusion once and serving three products is 3x more efficient than building three independent data stacks.
Derived attributes compound. Developer Trust Score computed for Broker also powers Analytix delay forensics. Cap rate benchmarks for Analytix also power Fractional AI overlay.
Ground truth flows back. Fractional's realised yields and vacancy data become validation inputs for Broker's yield projections and Analytix's cap-rate benchmarks. Real data beats modelled data.
Quality amortises. One quality framework, one conflict-resolution system, one lineage layer — applied consistently across all products.

How each product uses the shared layer¶

PropPie Fractional (AI overlay)¶

Already live. The AI layer adds:

Feature	Draws from
AI asset summary ("Why this, what could go wrong")	`ai.alpha_narrative`, `ai.risk_narrative`
Risk score on asset card	`risk.zone_risk_index`, `legal.title_clarity_score`
Portfolio fit indicator	`ai.persona_fit_score` (with consent)
Document Q&A on asset docs	RAG over asset documents
Distribution forecast	`ai.wealth_trajectory_paths`, `frac.realised_yield_ttm`
Comparable set	`ai.comparable_set`

PropPie Broker (B2C conversational)¶

New product. Draws most heavily from the data layer:

Flow	Draws from
Property lookup	`proj.`, `loc.`, `area.`, `party.`
Comparison mode	`ai.comparable_set`, all project attributes
Hidden costs	`ai.hidden_costs_breakdown` (stamp duty, GST, registration, society, etc.)
Title chain walkthrough	`legal.chain_of_title`, `ai.title_chain_explanation`
Wealth trajectory	`ai.wealth_trajectory_paths`, `ai.projected_irr_*`
Developer review	`dev.trust_score`, `ai.developer_track_record_summary`
GR/policy query	`policy.applicable_grs`, `ai.gr_impact_summary`
Market heat	`mkt.transaction_velocity_`, `mkt.median_price_per_sqft_`
"What could go wrong?"	`ai.brutal_honesty_flags`, `ai.risk_narrative`
Persona-aware ranking	`persona.*` → `ai.persona_fit_score`

PropPie Analytix (B2B dashboard)¶

New product. Uses aggregated and per-entity views:

Module	Draws from
Market Momentum	`mkt.transaction_velocity_*`, `mkt.sector_momentum_pct`, `mkt.price_appreciation_yoy_pct`
Delay Forensics	`proj.promised_completion_dates`, `proj.actual_completion_date`, `dev.trust_score`
Yield Heatmap	`mkt.cap_rate_median`, `mkt.yield_benchmark`, `loc.micromarket_*`
Comparable Analysis	`ai.comparable_set`, `fin.price_per_sqft_carpet`, `fin.asr_gap_pct`
Developer Intelligence	`dev.trust_score`, `ai.developer_track_record_summary`, `proj.estimated_cost_history`
Policy Feed	`policy.tailwind_flags`, `policy.headwind_flags`, `ai.gr_impact_summary`
Risk Radar	`risk.zone_risk_index`, `legal.title_clarity_score`, `risk.flood_zone_flag`
Asset Underwriting	All attributes on a single entity, packaged for due diligence

Cross-product data flows¶

The three products aren't just consumers — they contribute back:

flowchart LR
    FRAC[Fractional] -->|realised yields, vacancy, tenant data| Store[Canonical Store]
    BROKER[Broker] -->|user queries reveal demand patterns| Analytics[Usage Analytics]
    ANALYTIX[Analytix] -->|B2B user corrections, feedback| Store
    Analytics -->|popular micromarkets, trending queries| BROKER
    Analytics -->|feature usage, churn signals| ANALYTIX
    Store --> FRAC
    Store --> BROKER
    Store --> ANALYTIX

Key feedback loops: - Fractional → Store: realised yields on actual commercial assets validate the yield models used by Broker and Analytix. This is proprietary ground truth nobody else has. - Broker usage → prioritisation: the micromarkets and projects users ask about most drive pipeline prioritisation (scrape depth, freshness targets). - Analytix B2B corrections: enterprise users reporting errors improve data quality faster than automated QA alone.

Sequencing¶

Timeline	Product milestone	Data dependency
Now → Q3 2026	Data foundation v1	MahaRERA + IGR priority + GR classifier
Q3 2026	Broker private beta (100 users, Pune residential)	Full stack for Pune micromarkets
Q4 2026	Analytix design partner v0 (5 developers)	Market Momentum + Delay Forensics modules
Q4 2026	Fractional AI overlay v1	AI narratives + risk scores
Q1 2027	Broker public launch (Maharashtra)	Full stack MMR + Pune
Q2 2027	Analytix paid tier launch	All modules except Material Pulse
Q3 2027	Bilingual Broker (Hindi + Marathi)	Translation pipeline
Q4 2027	Karnataka pilot	New state data ingestion

Team allocation (to be refined)¶

Workstream	Lead	Developers
Data pipeline + ingestion	Vishal (CEO)	2 devs
PropPie Broker (B2C AI)	Aishvarya (COO)	1-2 devs
PropPie Analytix (B2B)	TBD	1 dev
PropPie Fractional AI overlay	TBD	Shared with Broker dev
Quality + ops	Shared	Shared

Total: 5 developers + 2 founders. Tight. The shared data foundation is the leverage — without it, this team size can't serve three products.

Revenue model interaction¶

Model element	Fractional	Broker	Analytix
Transaction / management fees	Yes	No	No
Subscription	No	Maybe	Yes
Freemium tier	No	Maybe	Maybe
Per-report / per-query	No	Maybe	No
White-label / embed	No	No	Yes
API access	No	No	Yes
Advertising / sponsored placement	Never	Never	Never

The "never advertise" commitment is structural: it's in the SOUL, the compliance docs, and this portfolio doc. The moment we take developer ad money, we lose the Honest Broker positioning.

What could go wrong at the portfolio level¶

Risk	Impact	Mitigation
Team spread too thin across three products	All three mediocre	Sequence ruthlessly; Broker beta before Analytix paid launch
Data foundation takes longer than expected	All products delayed	Vishal's pipeline is critical path; protect it
Fractional compliance event distracts from new products	Attention diverted	SM-REIT audit now; don't let it linger
Revenue pressure pushes us to take ad money	Moat destroyed	SOUL doc as institutional commitment; board-level rule
A product succeeds but cannibalises another	Revenue confusion	Clear ICP separation; Broker serves buyers, Analytix serves sellers/institutions

The one-sentence portfolio thesis¶

"One data foundation, three products, zero conflicts of interest — because the buyer is always the customer."

See also: - b2c-virtual-honest-broker.md — Broker full spec - b2b-realtyanalytix.md — Analytix full spec - proppie-fractional-context.md — Fractional as-built (placeholder) - wow-moments-catalog.md — Wow moments across all products - ../20-data/data-attributes.md — The ~140 attributes - ../10-strategy/vision.md — Strategic vision