The challenge
Skyline Estates started as two property professionals with a conviction: residential building management in South Africa runs on paper, WhatsApp groups, and month-end panic. They had validated the pain with 30 building managers, lined up three pilot customers, and had funding for roughly six months of runway. What they did not have was a product, or a technical team.
The brief was blunt. Get something real in front of the pilot customers within a quarter, prove people will pay, and do not build a throwaway. Plenty of MVPs win the demo and then collapse the first time a real portfolio lands on them. Skyline needed the opposite: a small product with a solid spine.
The approach
Our startup work follows a rule: cut scope, never quality. We ran a one-week discovery with the pilot customers and reduced the product to the three jobs a building manager does every single day. Everything else went on a list for later. Meanwhile we set up the boring things that make later possible: automated deployments, backups, monitoring, and a multi-tenant data model from day one.
- Cut the MVP to three jobs: levy billing, maintenance tickets, and resident communication
- Designed the data model multi-tenant from the first migration, so growth never forced a rewrite
- Set up CI/CD, monitoring, and automated backups in week one, before writing features
- Shipped to the three pilot buildings in week six and iterated on real complaints
- Charged from day one, because a pilot that pays is validation and a free one is a favour
What we built
The MVP was deliberately small: a building manager could bill levies, track maintenance, and message residents. The platform underneath was not small-minded. Every tenant lived in the same codebase behind clean isolation, every deploy was automated, and every error paged someone. That spine is why the same system that ran 3 buildings in week ten runs 200 today.
- Levy billing with automated statements, payment reconciliation, and arrears tracking
- A maintenance ticket flow connecting residents, managers, and contractors
- Resident notices and announcements that replaced the WhatsApp group chaos
- A multi-tenant architecture with per-building isolation and role-based access
- CI/CD through GitHub Actions: every merge deploys, every deploy is reversible
- Monitoring, alerting, and daily tested backups, in place before the first customer
6 weeks
to first pilot building live
3 to 200
buildings in 14 months, no rewrite
9 min
average time from merge to production
The outcome
Skyline signed its first paying customer ten weeks after kickoff and used that traction to close a seed round four months later. Fourteen months in, the platform manages 200 buildings and roughly 14,000 units, with uptime at 99.95% over the past year. Month-end billing, the moment the whole business depends on, has never been missed.
Just as important is what did not happen: no rebuild, no migration freeze, no quarter lost to paying down shortcuts. The team now ships several times a week through the same pipeline we set up in week one. Skyline has since hired two of its own engineers, who onboarded onto a documented, tested codebase, and we remain involved for architecture reviews and capacity planning.
