Keep identity behind the integration
My campaign workflows connect email clicks, page engagement, downloads and bookings with a pseudonymous identifier. The identifier can refer to a server-side record; it should not encode someone’s email address or expose that address in a URL. Pseudonymous data still needs access controls, an appropriate retention period and an agreed legal basis.
Model progress as separate events
A click, thirty seconds on a page, a PDF download and a booking represent different steps. I store them as distinct events or CRM properties rather than treating every click as a qualified enquiry. Brevo segments and attributes, HubSpot properties and timeline events, and n8n webhooks can carry this progression between tools.
Make retries safe
A webhook consumer should validate the incoming shape before updating a record. When a delivery is retried, an event identifier helps prevent the same operation being applied twice. I use explicit error paths and retry handling in automation workflows, with offline checks for Code nodes before deployment. Credentials stay in protected configuration.
Check what the evidence actually shows
Mail scanners can follow links before a person opens an email. I examine server logs and later engagement separately so those visits do not become misleading funnel milestones. For handover, I document the events, mappings and error paths. The useful result is a traceable workflow whose state can be checked at each stage, not a claim that every click came from a person.