Skip to content

Build Estimate — Housing Association Business Planning Platform

February 2026


The Core Thesis

Building a modern replacement for MRI HousingBrixx is a domain-complex but technically achievable project. The critical insight is this:

The financial modelling domain expertise — how housing associations model their 30-year plans, the accounting treatments, regulatory rules, covenant definitions, and stress test methodologies — is the most valuable and hardest-to-replicate input. The software engineering is the commodity part.

If the domain expertise comes from an advisor who already knows how every HA in the country builds their business plan, the build becomes a software engineering exercise around well-understood requirements. This dramatically reduces cost, risk, and time-to-market compared to a team that must learn the domain from scratch.


What Makes This Different from a Typical SaaS Build

This is not a CRUD application. The core challenges are:

  1. 30-year integrated 3-way engine correctness — P&L ↔ Balance Sheet ↔ Cash Flow with monthly/annual granularity over 360 periods, all deterministically linked
  2. Regulatory compliance outputs — FFR template with 132 financial lines + 160 assumption lines, SORP/FRS 102 conventions, RSH viability expectations
  3. Treasury/covenant logic under stress — configurable covenant definitions (varying by lender), headroom analysis, breach detection across multiple simultaneous scenarios
  4. Trust parity with incumbent — housing association treasurers will verify every number against their existing models. One rounding error = lost credibility

However, with the domain expert providing the financial logic, formulas, and accounting rules, the engineering team can focus on what they do best: building reliable, scalable, well-designed software.


Architecture Overview

┌─────────────────────────────────────────────────────────────┐
│                      Client Applications                     │
│  Modern Web App (React/Next.js)  │  Board/Read-Only Portal  │
└──────────────────┬──────────────────────────────┬────────────┘
                   │                              │
                   ▼                              ▼
        ┌──────────────────┐           ┌──────────────────────┐
        │  API Layer        │           │  Auth & Access        │
        │  (REST/GraphQL)   │           │  (SSO, RBAC, Audit)  │
        └────────┬─────────┘           └──────────┬───────────┘
                 │                                │
                 ├────────────────────────────────┘
                 ▼
      ┌──────────────────────────────────────────────────────┐
      │          Core Financial Engine                        │
      │  • Component compiler (building blocks)               │
      │  • Time-series calculation graph (360 months)         │
      │  • 3-way linkage (P&L, BS, CF)                       │
      │  • Covenant monitoring & breach detection              │
      │  • Scenario branching & comparison                     │
      └──────┬────────────────────────────┬──────────────────┘
             │                            │
             ▼                            ▼
  ┌────────────────────────┐   ┌──────────────────────────────┐
  │ Regulatory Reporting    │   │  Specialist Engines           │
  │ FFR export, SORP maps,  │   │  Treasury/loans, development  │
  │ VfM metrics, board packs│   │  appraisals, asset lifecycle  │
  └──────────┬─────────────┘   └──────────────┬───────────────┘
             │                                │
             └────────────┬───────────────────┘
                          ▼
           ┌──────────────────────────────────┐
           │  Data Layer                       │
           │  PostgreSQL │ Object Storage      │
           │  Redis/Queue │ Optional Warehouse  │
           └──────────────────────────────────┘
                          │
                          ▼
                ┌──────────────────────┐
                │  AI Layer (Phase 2+)  │
                │  Narratives, anomaly  │
                │  detection, drafting  │
                └──────────────────────┘

Key Technical Decisions

Decision Recommendation Rationale
Calculation engine Bespoke, deterministic Core domain moat; must be fully auditable and traceable. Cannot rely on third-party engine for regulatory-grade outputs
Frontend framework React/Next.js Large talent pool, strong ecosystem, server-side rendering for performance
Backend Python (FastAPI) or TypeScript (Node) Either works; choose based on team strengths. Python preferred for AI integration
Database PostgreSQL Reliable, well-understood, excellent for financial data; managed cloud options available
Hosting AWS or Azure UK regions UK data residency required for sector procurement; ISO 27001 achievable
Authentication Auth0 or similar managed service Buy, don't build. SSO/SAML support needed for enterprise clients
FFR export Custom Excel generation (openpyxl or similar) Must match NROSH+ Bulk Import Template exactly
Charting/dashboards Modern charting library (e.g., Recharts, D3) Board-quality visual output is a differentiator vs incumbent
AI/LLM API-based (GPT-4/Claude) Assistive only; never authoritative for compliant numbers

Build vs Buy Components

Component Decision Notes
Calculation engine Build Core IP and domain moat
Component framework Build Sector-specific building blocks
FFR export engine Build Must exactly match RSH template
Scenario manager Build Core differentiator
Authentication Buy (Auth0, Clerk) Commodity
Email/notifications Buy (SendGrid, Postmark) Commodity
File storage Buy (S3/Azure Blob) Commodity
PDF generation Buy (Puppeteer, WeasyPrint) Commodity
Hosting/infrastructure Buy (AWS/Azure managed services) Commodity
Monitoring/logging Buy (Datadog, Sentry) Commodity
AI/LLM capabilities Buy (API access) Use best-available models

Modules

Module Complexity Description
Core Financial Engine Very High 30-year 3-way integrated model, dependency graph, monthly/annual rollups, audit trace per number
Component System High ~100+ building block templates (tenures, loans, costs, development), formula-locked outputs
Scenario Manager High Copy-on-write branching, assumption overrides, multi-scenario comparison
Loan & Treasury Very High Multiple loan types, covenant calculation (configurable per lender), headroom analysis, breach detection
Regulatory Reporting Very High FFR export (exact template match), VfM metrics, SORP-aligned outputs, validation checks
Consolidation High Multi-entity consolidation with intercompany eliminations
Development Appraisal High Scheme-level modelling: land, build costs, tenure mix, grant, cross-subsidy
Asset Lifecycle High Component accounting, stock condition, decent homes, building safety, decarbonisation
Dashboard & Reporting Medium-High Board-ready dashboards, custom reports, PDF/Excel export, branding
Data Import/Export High Excel/CSV import, HMS/finance system connectors, migration tooling
User Management Medium Multi-tenant, RBAC, audit trail, SSO
AI Layer Medium Scenario narratives, anomaly detection, board paper drafting
Frontend Application Very High Planner UX, interactive modelling, scenario comparison, responsive design

Phased Build Plan

Phase 1 — Lean MVP (6–9 months)

Goal: Deliver a credible, auditable alternative for 3–5 pilot housing associations, with the advisor providing hands-on domain guidance throughout.

Team: Advisor (part-time domain lead) + 2–3 senior developers + 1 designer (part-time)

Includes:

Capability Detail
Core 3-way engine Monthly + annual rollups, 30-year horizon
Initial component library Income (by tenure), operating costs, capital programme, loans, equity
Scenario manager Base case + 3–5 stress scenarios
Loan/covenant essentials ICR, gearing, liquidity headroom (configurable definitions)
FFR export v1 SOCI, SOFP, cash flow — ready for NROSH+ upload
Basic reporting Board summary, covenant dashboard, VfM metrics
Import/export Excel/CSV templates for data onboarding
User management Tenant/user roles + full audit trail
Hosting Cloud-native deployment on AWS/Azure UK

Defers: Deep connectors, advanced consolidation, development appraisals, AI features, asset lifecycle depth, benchmarking.

Target cost: £100,000–£200,000

This assumes:

  • 2–3 senior developers at £500–£800/day (mix of employed and contract)
  • Part-time designer and part-time product/domain input from the advisor
  • AI-assisted development (Copilot, Claude, Cursor) accelerating code output by 30–50%
  • Modern frameworks reducing boilerplate dramatically vs 10 years ago
  • Cloud infrastructure costs minimal at pilot scale (<£500/month)
  • The advisor provides the financial modelling logic, formulas, and test data — this is the single biggest cost reduction vs a team learning the domain from scratch

Key milestone: Successfully produce an FFR-compliant output that reconciles against a pilot client's existing HousingBrixx output.

Phase 2 — Production Platform (+6–9 months)

Goal: Harden for 20–50 clients with reduced advisor involvement per implementation.

Team: +2–3 additional engineers, +1 QA, +0.5 SRE

Adds:

Capability Detail
Asset lifecycle module Decent homes, building safety programmes, EPC-driven decarbonisation
Consolidation v1 Multi-entity groups with intercompany eliminations
Development appraisal Scheme pipeline, grant assumptions, tenure mix, cross-subsidy
Richer dashboards Custom reports, PDF board packs, branded output
Integration layer Xero, selected HMS/GL imports
Migration toolkit Import adapters for HousingBrixx and common Excel model patterns
Compliance hardening Regression test harness against known FFR examples
SRE maturity SLAs, disaster recovery, monitoring, support workflows

Target cost: £200,000–£400,000 additional

Key milestone: Three clients successfully submit their FFR using the platform without parallel running against HousingBrixx.

Phase 3 — Scale & Differentiation (+12 months)

Goal: 100–200+ clients, category leadership, defensible moat.

Team: +3–5 additional engineers, +1 ML engineer, +1 data engineer

Adds:

Capability Detail
AI copilot Scenario narrative generation, anomaly detection, board paper drafting
Benchmarking Anonymised sector benchmarking across client base (opt-in)
Advanced treasury Hedging instruments, refinancing optimisation, debt strategy modelling
API ecosystem Open APIs, partner integrations, connector marketplace
Performance at scale Large portfolios (50K+ units), concurrent scenario runs
Advisor operating layer Multi-client command centre for advisory firms
International expansion Scottish/Welsh/NI regulatory variants; potentially Irish/Australian markets

Target cost: £400,000–£1,000,000 additional


Total Investment Summary

Phase Duration Team Size Cost Range Cumulative
Phase 1 — Lean MVP 6–9 months 3–4 people £100K–£200K £100K–£200K
Phase 2 — Production +6–9 months 6–8 people £200K–£400K £300K–£600K
Phase 3 — Scale +12 months 10–15 people £400K–£1M £700K–£1.6M

How This Compares

Approach Estimated Cost Time to Market
Traditional enterprise build (no domain partner) £3M–£8M 30–42 months
Advisor-led build (domain expertise provided) £700K–£1.6M 24–30 months
Lean MVP only (prove concept, iterate) £100K–£200K 6–9 months

The 3–5× cost reduction comes from three factors:

  1. Domain expertise is provided, not learned — the advisor knows exactly what the financial engine needs to calculate and why. No months of discovery, no expensive regulatory consultants, no getting it wrong and rebuilding
  2. Modern development tools — AI-assisted coding, modern frameworks, cloud-native infrastructure, managed services. A senior developer in 2026 produces 3–5× the output of a developer in 2015
  3. Focused scope — building only what housing associations need, not a general-purpose FP&A platform. Every feature serves the specific use case

What AI Can Automate

AI capabilities represent a genuine differentiator, not a gimmick. Housing association finance teams spend enormous time on tasks that AI can accelerate:

Capability Value Phase
Scenario narrative generation Auto-generate plain-English explanations of what each stress scenario means and why it matters — saving hours of board paper writing Phase 2–3
Anomaly detection Flag unusual assumptions or data patterns (e.g., "your void rate assumption is 2× the sector average for your region") Phase 2
Board paper drafting Generate first-draft business plan commentary from model outputs — the advisor reviews and refines rather than writing from scratch Phase 3
Natural language querying "What's our peak debt position in the combined stress scenario?" — instant answer from the model Phase 2
Automated stress test design AI generates plausible stress scenarios based on current economic conditions and RSH guidance Phase 3
Peer benchmarking insights "Your cost per unit is 15% above the sector median for associations of your size — here's what's driving it" Phase 3
Regulatory change impact When RSH updates the FFR template, AI helps map changes to existing models Phase 3

Important Principle

AI is assistive only — it never produces authoritative numbers for regulatory submissions. All financial calculations remain deterministic and fully auditable. AI helps with interpretation, narrative, and productivity, not arithmetic.


The Advisor's Role in the Build

The advisor is not a traditional "subject matter expert" hired for workshops. They are the domain co-founder — the person whose expertise makes the entire venture viable.

What the Advisor Provides

Input Why It's Critical Alternative Cost
Chart of accounts mapping Which building blocks map to which FFR lines, SORP categories, and covenant calculations 3–6 months of regulatory research + trial and error
Financial modelling logic How rent, costs, development, treasury, and depreciation interact over 30 years Would require hiring 2–3 specialist consultants at £1K+/day
Test data and validation Real (anonymised) business plans to test against — the only way to prove the engine works Cannot be synthesised; must come from real HA experience
Regulatory interpretation What the RSH actually expects vs what the rules literally say Years of sector experience
Covenant definitions How different lenders define ICR, gearing, asset cover — and the edge cases Proprietary knowledge from lender negotiations
Client relationships Warm introductions to 3–5 pilot clients willing to test 6–12 month cold sales cycle otherwise
Credibility A trusted advisor recommending the tool carries more weight than any marketing campaign Priceless

What the Advisor Doesn't Need to Do

  • Write code
  • Design interfaces
  • Manage cloud infrastructure
  • Handle customer support at scale
  • Build the sales and marketing machine (beyond initial pilots)

The advisor provides the what and why. The engineering team provides the how.


Risks and Mitigations

Risk Likelihood Impact Mitigation
Financial engine inaccuracy Medium Critical Golden test packs from real HA models; parallel running against HousingBrixx; formal validation with sign-off before each pilot goes live
FFR compliance gaps Medium High Embed regulatory SME (the advisor) in product loop; regression test harness against known FFR submissions; annual update process
Data migration friction High High Import adapters for common Excel patterns and HousingBrixx exports; assisted reconciliation (difference checker by statement/period); dual-run period tooling
Long sales cycles High Medium Advisor-led distribution bypasses procurement; start with 3–5 lighthouse pilots from advisor's portfolio; publish quantified outcomes
Incumbent lock-in High Medium Time market entry to coincide with SORP 2026 / rent settlement model refreshes; position as "run alongside" not "rip and replace" initially
Regulatory changes Ongoing Medium Annual FFR template mapping updates; advisor monitors regulatory pipeline; flexible engine designed for rule changes
Advisor availability Medium High Formalise advisor commitment early; document domain knowledge systematically; build internal capability over time
Security/procurement barriers Medium Medium ISO 27001 certification path from day one; UK data residency; standard HA procurement questionnaire readiness
Performance at scale Low (Phase 1) High (Phase 3) Architecture designed for horizontal scaling; performance testing with large portfolios from Phase 2

Ongoing Costs (Post-Launch)

Category Monthly Range Notes
Cloud infrastructure £500–£5,000 Scales with client count; managed services reduce ops overhead
Support & customer success £5,000–£15,000 Part-time initially, growing with client base
Continuous development £10,000–£30,000 Ongoing feature development and regulatory updates
Annual FFR template update £5,000–£10,000 One-off annual exercise (typically March, for June deadline)
Security & compliance £2,000–£5,000 Penetration testing, ISO 27001 maintenance
Total monthly burn £17,500–£65,000 Scales with maturity and client count

Revenue Required for Sustainability

At an average annual contract value of £15,000–£20,000:

  • Break-even on ongoing costs: 15–50 clients
  • Break-even on total investment (Phase 1–2): 25–60 clients
  • Profitable at scale: 100+ clients

Given the TAM of ~227 core buyers and the advisor's existing relationships, reaching 50 clients within 3 years of launch is a realistic target.


Timeline to Market

Milestone Target Date Assumption
Partnership agreement signed Q2 2026
Domain blueprint + chart of accounts mapping Q2–Q3 2026 Advisor-led, 6–8 weeks
Core engine + component framework + first UI Q3–Q4 2026 2–3 senior devs
Scenarios + covenant monitoring + FFR v1 Q4 2026–Q1 2027
Pilot onboarding (3–5 clients) Q1 2027 From advisor's portfolio
Parallel testing against 2027 FFR cycle Q1–Q2 2027 June 2027 FFR deadline
Production hardening + Phase 2 features Q3 2027–Q1 2028 Growing team
Scale to 20–50 clients 2028 Advisor-led distribution
AI features + benchmarking + 100+ clients 2029 Category leadership

Summary

A modern replacement for MRI HousingBrixx is technically feasible, commercially viable, and strategically well-timed. The key enabler is the advisor partnership:

  • Without domain expertise: This is a £3M–£8M, 3-year enterprise software project with high regulatory risk and uncertain market fit
  • With domain expertise: This is a £100K–£200K MVP that can be in pilot within 9 months, scaling to a £700K–£1.6M platform serving 100+ clients within 3 years

The software engineering is genuinely the commodity input. Modern tools, cloud infrastructure, and AI-assisted development mean that a small team of senior engineers can build in months what would have taken years a decade ago.

The domain expertise — knowing exactly what 227 housing associations need their 30-year business plan to produce, how every covenant is calculated, what the regulator actually expects, and having the relationships to get pilots through the door — that is the irreplaceable asset.


This document is part of a four-part research package. See also: Executive Summary · Product & Competitive Intelligence · Regulatory Requirements