Two-Way Sync and Multi-System Integration

When do you need two-way sync?

A one-way integration has a source and a destination. A two-way sync has neither, because the same customer can be edited in the CRM by a salesperson, in the accounting system by the bookkeeper, and on the website by the customer, all in the same afternoon. Multi-system work adds a third or fourth party: the job management platform that raises the invoice, the booking system that created the customer, the messaging platform that needs the current phone number. When more than one system can change a record and every system needs the current version, you are in this tier.

Typical shapes are a CRM such as HubSpot, Salesforce or Pipedrive kept in step with Xero, MYOB or QuickBooks; a field service platform such as ServiceM8, Tradify, simPRO or AroFlo synced with accounting and a customer-facing website; or a booking system such as Cliniko or Halaxy feeding payments through Stripe or Square and messages through Twilio, with the outcome written back. The standard integration build covers the one-way version of any of these, and scoping decides which tier a workflow belongs in.

Every build in this tier is priced from integration discovery and scoping. The data map and risk register from scoping are the inputs to the sync design, and the $1,800 + GST scoping fee is credited against the build when it proceeds within 30 days.

How is conflict handling designed?

Conflict handling starts before any code is written, with a decision per field about which system wins. Some fields have an obvious source of truth: the accounting system owns the invoice total, the booking system owns the appointment time. Others need a rule: most recent edit wins, one system always wins, or the conflict is held for a person to resolve. Those rules are written into the sync design document and agreed with you, because a sync that silently overwrites the bookkeeper's correction with a stale website value is worse than no sync at all.

The integration then enforces those rules, records every conflict it resolves in a log you can review, and holds anything it is not allowed to decide in a queue with a notification. Identity matching is designed the same way. Email addresses are not reliable keys, so the design names how a customer in one system is matched to the same customer in another and what happens when the match is ambiguous.

What does reconciliation do?

Sync failures are rarely loud. A webhook is missed during a vendor outage, a record is created while a token is expired, and the two systems drift apart by a handful of records a week that nobody notices until a customer is invoiced twice. Reconciliation is the scheduled job that compares the systems directly, independently of the event stream, and reports every mismatch it finds. Where the mismatch has a rule, it is repaired; where it does not, it is listed for review. The reconciliation report is the evidence that the systems agree, and it runs whether or not anything has changed.

What testing is included?

The full integration assurance suite is bundled, never an optional line. Contract tests run against every vendor API in CI so a change on the vendor's side fails a test before it corrupts data. End-to-end tests push real records through the real workflow in each vendor's sandbox in both directions. Rate-limit tests run against each vendor's documented caps, because a multi-system sync multiplies calls. Failure-mode tests cover token expiry, webhook retry, partial failure and duplicate delivery, and prove the sync is idempotent in every direction. Security checks confirm credentials are backend-only and that no endpoint exposes one customer's records to another. Monitoring and alerting on sync health and reconciliation results are wired into the CI pipeline.

How is it priced and what do you own?

The price is from $8,000 + GST, fixed for your build in the scoping quote. The number of systems, the number of fields with real conflict rules and the volume the sync must sustain decide the figure. Third-party platform and API fees are named in the quote, and who pays each is settled in writing before work starts. A sync that touches Xero is designed around webhooks and delta sync rather than polling, because Xero now charges for API access on usage-based tiers, and that fee is named in the quote.

You own the code, the repository, the vendor accounts, the credentials, the test suites, the monitoring and the runbook. Any developer can maintain it. The integration support retainer is available month to month, with no lock-in, if you would rather we watch it.

What if you need a portal or dashboard?

The moment the work needs a logged-in area for customers, a dashboard that shows data from several systems, or a bespoke workflow that none of the existing platforms provides, it is a web app or client portal, priced from $12,000 + GST. It is a different kind of build with its own scoping, and that page describes it. Scoping tells you which side of that line your work sits on.