Your integration failed. Who finds out?

When a donation moves from a fundraising platform into a CRM, there should be a clear way to confirm that it arrived, was recorded correctly and is ready for the next step. Good integration design builds that assurance into the connection from the outset.

For a charity, that means confidence that supporters receive the right communications, Finance can reconcile income, and fundraisers can work from reliable records. The controls that support those outcomes deserve as much attention as the connection itself.

Define what “working” means

Start by agreeing what a successful transaction looks like. For an online donation, that might mean the payment reference is stored in the CRM, the gift is linked to the correct supporter and campaign, any Gift Aid declaration is captured, and the record is available for acknowledgement.

Those criteria give you something meaningful to monitor. Alongside checking that the integration is running, you can check that transactions reach their destination and contain the information needed by the teams using them.

An operational owner should help define these expectations and oversee the integration in day-to-day use. They understand which information matters, how quickly it needs to arrive and when an issue needs attention. Working with the technical team, they can agree alert thresholds, response times and escalation routes that fit the charity’s operations.

Build in traceability, reconciliation and safe recovery

Each transaction should carry a consistent reference through the connected systems. That gives support teams a practical way to trace a donation from the original payment to the CRM record and any subsequent processing.

Reconciliation adds another layer of assurance. Comparing transaction counts and values between systems, with agreed allowances for timing differences, helps confirm that the complete set of expected donations has arrived. It can also identify gaps that individual error alerts would miss.

Recovery should be part of the design too. Temporary connection problems may suit automatic retries. Records requiring a correction, such as a missing campaign mapping, need a clear route for review and replay. Using the original transaction reference to recognise records already processed helps make those retries safe and prevents duplicate gifts.

Together, these controls give teams a way to confirm completeness, investigate individual records and resolve exceptions with confidence.

Make the controls easy to operate

Useful monitoring tells the right person what needs attention and gives them enough information to act. An alert should identify the transaction, source system, timestamp and reason for the exception, with a clear route to the relevant record or processing log.

A short runbook makes that process repeatable. It should explain the normal flow, where monitoring lives, how common exceptions are handled, who provides cover and how recovery is confirmed. Someone taking over support should be able to use it without needing the original developer beside them.

Before go-live, test these controls alongside the normal transaction flow. Demonstrate that a donation can be traced, reconciled and safely replayed, and that the people responsible know how to respond.

That is what makes an integration dependable: clear evidence that information has arrived correctly, supported by a practical process for keeping it that way.