Multi-Tenant SaaS Platform for Har***** Property, AustraliaCase Study

The existing application worked well for one agency and made assumptions everywhere that only one agency existed. Turning that into SaaS is not a configuration change; it is an architectural rebuild, and the client had underestimated that. Tenant isolation was the hard requirement — property management involves tenant personal data, lease documents and trust-account-adjacent financial records, and any leakage between agencies would be an existential problem for a young product. Australian property management operates under state-based residential tenancy legislation with differing notice periods, bond handling and entry rules, so a national product cannot hardcode one state's rules. Trust accounting obligations meant the financial side had to be handled conservatively, with the clear position that the platform records and reconciles but does not become the trust account system of record. Document storage volume was significant, with lease documents and condition reports including large photo sets. There was also commercial pressure: the client had letters of intent from early customers contingent on delivery within a defined window.

We rebuilt on Next.js with a Node.js API and PostgreSQL, implementing tenant isolation at the data layer with row-level enforcement rather than relying on application-level filtering — application filters fail open when someone forgets a where clause, and that failure mode was unacceptable here. Every query path was reviewed against isolation, and we built an automated test suite that specifically attempts cross-tenant access on each endpoint, run on every deploy. State-specific tenancy rules are configuration rather than code, so adding a state is a data exercise. Documents and condition report photos go to S3 with per-tenant prefixes and signed time-limited URLs, never public objects. Stripe handles subscription billing with per-tenant plans and seat counts. Financial records are structured to reconcile against the agency's trust account with a clear boundary — the platform is deliberately not positioned as the trust account itself, which we advised on directly given the regulatory exposure. Onboarding was built as a self-service flow because a SaaS product requiring manual setup per customer does not scale. We also built tenant-level backup and export, so an agency can leave with its own data — a question both launch customers asked during due diligence, and one a new SaaS product needs a credible answer to before it will be trusted with a portfolio. Related: custom software development and our real estate practice.