Back to Case Studies

Reference Architecture

Architecture pattern — built and tested. Not currently in production with paying customers. See wintura.ai for Soatech's flagship production case study.

Real EstateWeb DevelopmentCloud & DevOpsIteration Sprint

PropTech Platform: The Iteration Sprint Scaling Pattern

Architecture pattern — built as a reference implementation, not in production with paying users. The featured production case study is wintura.ai (see /case-studies/wintura-ai). This pattern documents the PropTech Iteration Sprint playbook a Soatech engagement would follow for that domain.

Reference Implementation6 months1 team members
Velocity Proof

Architect-led, AI-accelerated. 2× faster than hand-coded shops.

Traditional Agency
12–14 months

Hand-coded teams, hourly billing, scope creep. Multi-month ramps before the first production-grade PR.

Soatech + AI — Architect-Led
6 months

A Veteran Architect leads the Pod. AI tooling captured as reviewed throughput, not someone else's margin. Fixed sprint price.

Same scope, same quality bar. the speed.Pre-Blueprint engagement — outcome shown for reference

Key Results

0x
Feature Velocity
Monthly feature releases vs the prior cadence
0%
Infrastructure Costs
Reduction after cloud optimization
0%
Response Time
Faster average API response after tuning

The Challenge

This reference implementation documents the Iteration Sprint pattern for a property management SaaS that has outgrown its prototype. The scenario: a London-based platform with 500+ landlord customers and 8,000+ rental units, where the existing codebase ships features too slowly to keep up with demand.

  • Tenant portal is the #1 requested feature — tenants need to pay rent, submit maintenance requests, and view documents online
  • Maintenance workflows are manual — property managers coordinate repairs via email and phone
  • Infrastructure costs are ballooning — the AWS bill has doubled in 6 months without corresponding traffic growth
  • Technical debt from early-stage shortcuts slows every new feature

The constraint is delivery throughput, not headcount. The Iteration Sprint model answers it differently than staff augmentation: instead of adding bodies, a senior architect owns the roadmap end-to-end and ships scoped feature modules every two weeks at a fixed price per sprint.

The Approach

Sprint cadence: fixed-price feature modules

Under the Iteration Sprint model, your architect delivers one to three feature modules per two-week sprint. Each module is scope-capped (≤5 screens, ≤3 entities, ≤2 integrations) so the price is fixed and the timeline is predictable. There is no day-rate, no hourly billing, and no open-ended "team" you have to manage.

The working rhythm for this pattern:

  • Shared access to Slack, Linear, and GitHub so reviews and demos happen in the open
  • A bi-weekly sprint demo plus async updates between demos
  • 11:00 CET overlap with London (1 hour ahead) so feedback loops stay tight
  • Every PR reviewed before merge — architectural coherence is owned by one person, not diffused across a rotating roster

Sprints 1–4: Tenant Portal

The first feature module shipped was the tenant portal:

  • Rent payments — Stripe integration with automated reminders and receipt generation
  • Maintenance requests — Photo upload, categorization, and real-time status tracking
  • Document vault — Tenancy agreements, inspection reports, and correspondence in one place
  • Mobile-responsive — 65% of tenants access the portal on mobile

The portal pattern is designed to reach activation quickly: in this reference scenario, 72% of tenants activated their accounts within the first month.

Sprints 5–8: Maintenance Automation

The highest-impact module replaces a chaotic email-based workflow with an automated system:

  • Smart routing — Maintenance requests automatically categorized and assigned to the right contractor based on type, location, and availability
  • Contractor portal — External tradespeople accept jobs, update status, and upload completion photos
  • SLA tracking — Automatic escalation if repairs aren't acknowledged within 4 hours or completed within the agreed timeframe
  • Cost tracking — Every maintenance job linked to the property with full cost history

In this reference scenario, property managers save 3–4 hours per day on maintenance coordination.

Ongoing: Infrastructure Optimization

Infrastructure work runs as an Iteration Sprint Plus module (feature modules plus ongoing CI/CD, observability, and performance tuning):

  • Right-sized EC2 instances — Most were over-provisioned by 2–3x
  • Reserved instances — 40% savings on compute costs
  • Caching layer — Redis for frequently-accessed property and tenant data
  • Database optimization — Query analysis and indexing reduced average response time from 800ms to 120ms
  • CI/CD pipeline — Automated testing and deployment replaced manual processes

In this reference scenario, the monthly AWS bill drops from £8,200 to £4,900.

The Results

After 6 months of bi-weekly sprints, the target outcomes for this pattern:

  • Feature releases: From 1–2 per month to 5–6 per month
  • Tenant portal adoption: 72% of tenants active within 30 days of launch
  • Maintenance efficiency: 15+ hours per week saved on coordination
  • Infrastructure: 40% cost reduction, 85% faster average response time

The same architect carries context across every sprint, so the codebase stays coherent as it grows — and the engagement converts naturally into an ongoing Iteration Sprint or Production Subscription once the roadmap stabilizes.

Why It Works

  1. Timezone alignment is underrated. Soatech operates CET — just 1 hour ahead of London. Demos, reviews, and Slack conversations happen naturally, not at awkward hours.

  2. One architect, continuous context. A single senior architect owns the roadmap across every sprint. Architectural decisions compound instead of fragmenting across a rotating set of contractors.

  3. Fixed-price discipline. Each feature module is scope-capped and priced before work starts. There is no hourly meter and no scope creep — out-of-scope requests roll into the next sprint, never silently into the current one.

Technology Stack

Next.jsNode.jsPostgreSQLRedisAWSTerraformDockerGitHub Actions

Want us to build something like this?

Let's discuss how we can help your business grow.

Get in Touch
Book a Production Audit · €1,500