Context
Raketech is a publicly listed performance marketing company operating a portfolio of 8 products across Asia, Europe, and the US. Traffic and content were strong. What happened after the first visit was not. There was no CRM function: no lifecycle campaigns, no segmentation, and no way to know which messages moved revenue.
The problem
Starting from zero is a different problem from optimising an existing function. There was no campaign taxonomy, no audience model, no attribution framework, and no operating rhythm. On top of that, whatever I built had to work across 8 products in different markets and languages from day one.
What I did
Designed the campaign taxonomy first: a naming and structure standard that made every campaign measurable and comparable across all 8 products before a single message went out.
Built the audience segmentation model on behavioural and lifecycle data, so campaigns targeted states (new, activating, at-risk, dormant) rather than lists.
Stood up the delivery engine in Braze and scaled it to 6+ campaigns per week across acquisition, activation, and retention.
Established the attribution framework in GA4, so CRM impact was measured on conversions rather than opens.
Built real-time KPI dashboards in Looker Studio on top of GA4 and campaign data, so product leads could see conversion, retention, and campaign performance themselves instead of asking me for a number.
Ran structured test-and-learn cycles with VWO. Every campaign was a hypothesis, and every result went back into the playbook.
The outcome
The function outlived my tenure. That is the real test. Taxonomy, segmentation, dashboards, and playbooks were all documented and handed over as a working capability rather than a dependency on me.
Need the same thing done? Here is how a CRM function build engagement works.