Connect or replace: a saner path off a legacy property stack

Every REIT we speak to runs some version of the same estate: an ageing property management system at the centre, a CRM bolted to the front, a finance system at the back, and a lattice of spreadsheets holding it together. Everyone agrees it should change. Nobody wants to be the executive who signed off the big-bang migration that stalled leasing for a quarter.

The hesitation is rational. The legacy stack touches everything - every lease, every debit order, every statement run. Replacing it in one movement means migrating live financial data, retraining every team, and cutting over collections in a single weekend, usually with the next billing run days away. When that goes wrong, it goes wrong in public: rent not collected, statements not sent, funder claims missed.

So portfolios sit still, and the cost of sitting still compounds quietly - more spreadsheets, more manual reconciliation, more decisions made on stale numbers. The trap is treating "replace everything" and "change nothing" as the only options. There is a third: connect first, replace deliberately.

In connect mode, the new platform syncs inventory, tenancies and balances from the systems you already run. Nothing is cut over. Sync workers pull on a schedule, map fields between the two worlds, flag conflicts instead of silently overwriting, and surface monitoring so you can see exactly what flowed and when. Your teams keep working where they always have.

What you gain immediately is the thing the legacy estate can never give you: one live, consolidated view of occupancy, collections and arrears across the whole portfolio. The board pack stops being a month-end assembly job. Ops and finance stop arguing about whose number is right, because there is finally one number, computed in one place, from the freshest data each source system holds.

Connect mode is also the cheapest possible due diligence. Before you migrate anything, you learn exactly how clean - or not - your legacy data is, because the sync surfaces every anomaly: leases with no linked tenant, balances that do not add up, units that exist in one system and not the other. You fix data quality while the legacy system is still the system of record, not during a cutover weekend.

This is only possible if the platform is API-first in both directions - able to ingest from your stack on a schedule with mapping and monitoring, and able to expose everything it holds back out over documented endpoints and webhooks. If the API is an afterthought, connect mode is a brochure feature.

There are reliable warning signs. If "integration" means a CSV export emailed weekly, that is not integration. If every new endpoint is a paid scoping exercise with a six-week lead time, the vendor is selling professional services, not a platform. If the API only covers a fraction of what the product does, you will hit the wall exactly where it hurts - usually somewhere in billing.

The test we hold ourselves to at Atrium is structural: every capability ships with its endpoint, its webhook event and its OpenAPI documentation, or it does not merge. The API is never behind the product, because it is not allowed to be. Whatever platform you choose, ask for the equivalent guarantee - and ask to see the API catalogue before you see the sales deck.

With the consolidated view running, replacement becomes a sequence of contained decisions instead of one existential one. Each move is scoped to a workload and a slice of the portfolio, has a rollback, and produces a measurable result before the next move starts.

A typical sequence: move collections for one property first, because that is where the pain is sharpest and the win most visible - debit orders, instant EFT and reconciliation running on the new platform while everything else stays put. Then run the next intake's applications end-to-end: applications, vetting, allocations and e-sign leases are naturally seasonal, so a new cycle can start clean on the new system without migrating historic workflow. Maintenance moves when the field team is ready, not before. Accounting moves last, at a financial year boundary, once months of parallel-run data have proven the ledgers agree.

Sequencing also respects the calendar. Residential portfolios have rhythms - academic intakes, billing runs, audit season - and each workload has a natural window where switching is low-risk. A big-bang migration has to fight every rhythm at once; a deliberate sequence surfs them.

It helps to name the actual surfaces, because "integration" hides five different jobs. Inventory sync: properties, units and beds flowing in with stable identifiers, so both worlds agree on what exists. Tenancy and balance sync: leases and account positions, the data with the least tolerance for drift. Money: payment providers and debit-order rails webhooking settlement events in near-real time, because collections cannot run on yesterday's file. Communications: email, SMS and WhatsApp providers, with delivery status flowing back onto the record. And the general ledger: summarised journals out to Xero or your GL on schedule, so accounting reflects operations without re-keying.

Each surface has its own rhythm and its own failure modes, which is why monitoring is not optional. A sync that fails silently is worse than no sync at all - it manufactures confident, wrong numbers. Every flow needs a visible last-run status, a diff of what changed, and an alert when something that should have arrived did not.

Ask any vendor four questions about this surface. Can I read everything out that I put in, without a services engagement? Do state changes emit events I can subscribe to, or must I poll? Is the API versioned, so your upgrades do not break my integrations? And does the permission model govern the API exactly as it governs the screens? The answers predict your migration experience better than any reference call.

Shorter than the big bang, despite appearances. The big-bang project spends months in workshops trying to de-risk a single irreversible weekend, and the risk never actually leaves. The connect-first path ships value in the first weeks - the consolidated view - and then converts risk into a series of small, reversible steps. Momentum is easier to sustain because every step pays for itself.

The endgame is the same: one platform as the source of truth, the legacy estate retired, the spreadsheet lattice dissolved. The difference is that you get there without ever betting the portfolio on a weekend - and you can stop, pause or reorder the journey at any point, because every intermediate state is a working state.

That is the whole argument. Do not choose between a dangerous leap and an expensive standstill. Connect first. Replace deliberately. Keep the rent flowing the entire time.