Full BA-backed platform
Not a $30k build
If the full 22-epic platform is required now, the funding model must grow beyond a simple low-cash custom build.
MHMW Financial Proposal
Price the journey in phases, not the entire destination at once
The BA work defines the full destination. The currently discussed budget defines what can realistically be funded first. This proposal translates that gap into delivery options.
Budget-fit first build
$25k-$39k
A disciplined first build can fit closer to the current budget if the team accepts phased expansion.
Best partnership use
Lower cash, clear upside
Partnership works only when revenue share, reporting, timing, and ownership are all clearly defined.
Funded MVP budget
$25k-$39k working envelope
This is the recommended funding view for the first meaningful build. It preserves the larger vision but prices only the first disciplined operational release, not the entire 22-epic platform.
Phase 0 — discovery and deal shaping
$3k-$6k
Used to confirm scope, decisions, architecture, and commercial structure.
Phase 1 — funded MVP build
$18k-$25k
Used to build the first meaningful operational core across the four MVP epics.
Phase 2 — stabilization and rollout support
$4k-$8k
Used to harden the first release enough for controlled use and launch support.
Phase pricing
Suggested first funding structure
This model assumes the client cannot fund the full platform now and needs an honest first-build path that still preserves the larger vision.
| Phase | Purpose | Duration | Estimate |
|---|---|---|---|
| Phase 0 | Discovery and commercial shaping Clarify scope, resolve decisions, confirm architecture, and lock the commercial structure. | 2-3 weeks | $3,000 - $6,000 |
| Phase 1 | Funded first build Deliver the highest-value operational core rather than the full 22-epic platform. | 6-10 weeks | $18,000 - $25,000 |
| Phase 2 | Launch support and stabilization Cover rollout support, fix-forward work, deployment hardening, and validation. | 2-4 weeks | $4,000 - $8,000 |
Epic breakdown
Backlog-backed first-build estimate
These ranges tie the financial proposal back to the MVP epic structure rather than treating the build as one undifferentiated number.
| Epic | Scope | Stories | Weight | Estimate |
|---|---|---|---|---|
| EPIC-MVP-001 | Secure Listing Foundation | 11 | 25 | $9.5k-$14.8k |
| EPIC-MVP-002 | Buyer Enquiry and Viewing | 6 | 13 | $5.75k-$8.97k |
| EPIC-MVP-003 | Disclosure, Offer and Handoff | 4 | 11 | $4.75k-$7.41k |
| EPIC-MVP-004 | Operations and Safety | 5 | 12 | $5k-$7.82k |
Budget fit
What the current budget can actually support
A budget around $30,000 can support a disciplined first build, but not the entire full-platform vision represented by all 22 epics.
If the client insists that the full scope is mandatory from day one, the commercial model must change through either higher funding, phased milestones, or a serious partnership structure.
The most honest near-term position is to preserve the full roadmap while funding only the first meaningful leg of the journey now.
Recommended target
$25k-$39k first build envelope
This range supports paid shaping, a smaller funded build, and a short stabilization phase without pretending that the full BA-backed platform has been paid for.
Explicit exclusions
What this funded MVP budget does not include
These exclusions protect the budget from quietly expanding into the full-platform scope.
Premium subscriptions and promoted listings
AI image enhancement and AI listing-description generation
Property reels, audio stories, and automated valuation
Marketing-material generation and QR workflows
Calendly dependency and external calendar write-back
Contract-pack generation and broad e-signature workflows
Escrow instructions, earnest money, and option-fee handling
Inspector marketplace, scheduling, payments, reports, and payouts
Post-inspection renegotiation and amendment workflows
Title review, issue management, and clearance workflows
Financing, appraisal, closing preparation, and final walkthrough
Funding, recording, transaction completion, and post-closing ratings
Infrastructure position
Why the self-hosted DIY path still makes business sense
This should be framed as a commercial choice, not a hobbyist preference. The managed cloud version mainly buys convenience and outsourced operations.
| Topic | Self-hosted DIY | Managed cloud |
|---|---|---|
| Setup and convenience | The team owns Docker, reverse proxy, SSL, and deployment discipline. | Cloud setup is faster because provisioning and public access are pre-managed. |
| Backups and recovery | MHMW must script backups, store them elsewhere, and test restores regularly. | Managed plans bundle backup and recovery tooling. |
| Scaling | Growth is handled through server sizing and deliberate architecture upgrades. | Managed plans usually make scaling easier to trigger operationally. |
| Monitoring and security operations | MHMW must own observability, patching, firewall rules, SSL lifecycle, and secret rotation. | A managed platform bundles more of the infrastructure operations surface. |
Why MHMW should still prefer self-hosting
Lower recurring platform markup once the initial setup discipline is in place.
Better control over property data growth, audit trails, storage, and operational boundaries.
Stronger long-term economics for a data-heavy real-estate platform.
Ability to colocate app, database, and future AI services for lower latency and tighter control.
The trade-off we must own
The savings are real only if MHMW also owns backup automation, restore drills, patching, observability, secret rotation, and deployment hardening. Self-hosting is cheaper than managed cloud only when operator discipline is treated as part of the product cost.
Equivalent infrastructure cost
What this table is actually showing
This is not a Supabase feature comparison. It is a raw infrastructure-cost illustration showing that if MHMW is willing to operate its own Linux server, managed cloud pricing includes convenience and operations markup on top of compute.
The point is simple: for broadly similar infrastructure capacity, self-hosted servers are often far cheaper than hyperscaler or fully managed platform pricing.
So when we choose self-hosted Supabase/Postgres, we are not claiming the managed product has no value. We are showing that the extra spend mainly pays for convenience, managed backups, easier scaling, bundled observability, and lower operator burden.
| Provider | Specs | Monthly cost |
|---|---|---|
| Contabo (Cloud VDS L) | 6 physical cores / 48 GB RAM / 360 GB NVMe | €64/mo (~$74) |
| Hetzner | 8 vCPU / 32 GB RAM / 240 GB NVMe | ~$50 |
| DigitalOcean | 8 vCPU / 32 GB RAM / 320 GB SSD | ~$168 |
| AWS EC2 (m5.2xlarge) | 8 vCPU / 32 GB RAM | ~$280 |
Contabo is included here deliberately because it reflects the owned-infrastructure posture already being considered for MHMW. In other words, this is not a hypothetical DIY path; it matches the operating model we are actually proposing.
Service-by-service savings
Where the self-hosting savings actually come from
The biggest savings do not come from regulated trust services. They come from product infrastructure that would otherwise carry recurring SaaS markup.
| Service area | Typical hosted posture | Self-host / hybrid posture | Indicative annual saving |
|---|---|---|---|
| Managed database + backend platform | Supabase Cloud, Neon, or RDS plus backend-platform markup | Self-hosted PostgreSQL + Supabase on Contabo/Hetzner-class infrastructure | $3,000-$9,000 |
| Object storage + media delivery markup | Supabase Storage Cloud, S3-only managed posture, or storage vendor markup | Supabase Storage or MinIO-style S3-compatible storage plus chosen CDN posture | $1,200-$4,800 |
| Monitoring and error tracking | Hosted Sentry-class tooling | Self-hosted Sentry or Prometheus + Grafana + Loki / OpenTelemetry stack | $600-$3,600 |
| Product analytics | Hosted PostHog-class event analytics SaaS | Self-hosted PostHog or equivalent | $1,200-$6,000 |
| Realtime / chat infrastructure | Managed chat or realtime vendor | App-owned realtime on Supabase Realtime / Postgres event infrastructure | $1,200-$4,800 |
| PDF / QR / document utility services | Multiple small SaaS utilities or workflow add-ons | In-app libraries, headless browser generation, and worker jobs | $300-$1,500 |
| Malware scanning | Managed file-security add-on | Internal ClamAV-style scanning worker | $600-$2,400 |
| Email in lower environments | Paid delivery across all environments with Resend/SMTP-class providers | Production SaaS only, self-managed dev/test and internal-only mail paths | $200-$900 |
| Estimated total annual saving | — | — | $8,300-$33,000 |
These savings do not remove Stripe, telecom delivery, identity verification, e-signature, or map/geocoding costs. Those remain mostly external even in a self-hosted product.
Buy vs own boundary
Which vendor categories should still stay external
The financial case is strongest when MHMW self-hosts product infrastructure but keeps regulated or hard-to-replicate trust rails external.
| Service boundary | Typical vendors | Why we still buy this |
|---|---|---|
| Payments and payouts | Stripe Payments, Stripe Connect | PCI scope, disputes, KYC, payout rails, and reconciliation are not worth rebuilding for the pilot. |
| E-signature | Zoho Sign | Certificate chains, signer audit, and legal trust are better bought than self-hosted early. |
| Identity and telecom | Identity-verification provider, Twilio Verify / SMS | Liveness, ID verification, and telecom delivery are compliance-heavy and operationally specialised. |
| Licensed property and map data | HCAD / county CAD sources, commercial geocoding/maps | Licensed data access, coverage governance, and map datasets remain better consumed externally. |
Property data coverage cost
HCAD is one county source, not the whole U.S. data strategy
If MHMW wants Texas-wide and then nationwide property coverage, that becomes a separate commercial and technical workstream rather than a small follow-on API task.
| Coverage layer | Likely sources | Commercial implication |
|---|---|---|
| Harris-only pilot | HCAD | Lowest-cost entry point, but only Harris County appraisal coverage. |
| Texas county expansion | Fort Bend CAD, Dallas CAD, Bexar Appraisal District, Travis CAD, Tarrant Appraisal District, Montgomery CAD, and other county sources | Adds ingestion, normalization, and county-specific maintenance work. |
| Texas listing / market expansion | HAR and other approved MLS / RESO Web API feeds | Introduces listing-data licensing, compliance, and display-use constraints. |
| Nationwide expansion | Regrid, ATTOM, Cotality/CoreLogic-class, DataTree/First American-class, or equivalent commercial aggregators | Usually introduces enterprise pricing, usage limits, and wider legal/commercial review. |
This is the key financial point: HCAD integration does not mean “U.S. property data is solved.”
Broader coverage may require county-by-county ingestion, MLS licensing, national aggregator contracts, geocoding usage, and legal review of data-use rights. That should be treated as an explicit budget line, not hidden inside general engineering effort.
Commercial options
Ways to bridge vision and affordability
Cash-only phased build
$25k-$39k total
Best when the client wants straightforward contracting and can fund the first build directly.
Use a paid discovery, a fixed first build, and a small stabilization phase. This is the cleanest commercial model.
Lower cash plus revenue share
$10k-$20k upfront + agreed revenue share
Best when the client cannot fully fund the build now but the venture has believable upside.
Reduce upfront cost, then recover value through a fixed-duration or capped revenue share arrangement.
Capped recoup model
$15k-$20k upfront + capped recovery target
Best when both sides want partnership but also want a commercial ceiling and clearer recovery path.
Use a reduced upfront fee and recover the remaining value until a pre-agreed cap is reached.