Every fractional CMO engagement I have seen fail shared one pattern: the external hire started by doing things rather than understanding the system. They launched campaigns before auditing attribution. They changed the messaging before understanding why conversions were dropping. The first 90 days are not an execution phase, they are a diagnostic and architecture phase. Execution without the right foundation just produces faster wrong results.
Weeks 1–2: Diagnostic
Before any campaign goes live or any budget shifts, I spend two weeks understanding what is actually happening in the business, not what the founders believe is happening, but what the data shows. The diagnostic covers four areas in sequence: attribution setup (where signal is leaking, what the CRM is and is not capturing), campaign architecture (how campaigns are structured, what conversion events they are optimising against, whether those events correlate with revenue), CRM data quality (which fields exist, which are populated, what the data actually contains), and GTM governance (who has access, what tags are firing, what has changed in the last six months). The output is a constraint map: the one or two things that are limiting the performance of everything else.
Weeks 3–6: Architecture
Once the constraint is identified, I fix the foundation before touching acquisition spend. If the constraint is signal, the first project is server-side CAPI and GA4 event cleanup. If the constraint is attribution, it is building the GTM → GA4 → CRM → Looker chain. If the constraint is GTM architecture, it is restructuring campaigns around correct conversion events and offer clarity. If the constraint is RevOps, it is fixing CRM fields, handoff processes, and the MQL definition. Weeks 3–6 produce infrastructure. The work is invisible from the outside, no new campaigns, no new creative. But it is the work that makes everything after it accurate and scalable.
Weeks 7–12: Execution
With signal clean and attribution trusted, acquisition is finally optimisable. Campaigns are restructured around the correct conversion events that actually correlate with revenue. Budget is reallocated based on attribution data rather than channel-reported ROAS. The operating cadence is established: weekly performance reviews with defined owners, monthly strategy reviews that connect spend to pipeline, anomaly thresholds that trigger action without a meeting. New campaigns are tested against a proper incrementality framework rather than blended ROAS. By week 12, the system should be running with enough momentum and clarity that the team can identify problems and escalate them without waiting for the fractional CMO to notice.
The handover is the deliverable
The engagement ends with a system that runs without me. That means documentation: attribution architecture decisions with the rationale for each choice, campaign structure with the logic behind the organisation, operating playbooks for weekly and monthly reviews, decision frameworks for common scenarios (CPL spike, creative fatigue, budget reallocation requests). The goal is that six months after the engagement ends, the team could explain to a new hire why every piece of the system works the way it does. Most fractional arrangements end when the work becomes comfortable for the external hire. The model I use ends when the system is autonomous.