A fractional CMO’s first 90 days should run on a written plan with deliverables at fixed gates, not on vibes and weekly syncs. Below is the week-by-week template I use; companies can hand it to any operator they hire and ask them to mark it up. The structure is four phases: diagnose, decide, build, scale, each ending in something concrete you can inspect.
I’m Andrii Byzov, an AI-native fractional CMO for B2B tech. This is the actual shape of my engagements, with the client-specific parts replaced by brackets.
Key takeaways
- Four phases with gates: diagnosis (weeks 1-3), strategy sign-off (weeks 4-5), system build with live experiments (weeks 6-9), scale and renewal basis (weeks 10-13).
- Every phase ends in an artifact: an audit document, a one-page strategy, a live dashboard, a renewal memo. Activity without artifacts is how engagements drift.
- Expect diagnosis before tactics. An operator proposing channel spend in week 1 is pattern-matching from their last client, not learning your business.
- The 90-day question that matters at the end: would the system keep working if the operator left tomorrow?
The template
Phase 1: Diagnose (weeks 1-3)
Week 1: access and baseline.
- Access to analytics, CRM, ad accounts, content repositories, prior strategy docs.
- Baseline metrics pulled: traffic, conversion points, pipeline by source, CAC if calculable, sales cycle length, win rate.
- Leadership interviews: CEO, sales lead, product lead. [30-45 min each]
- Deliverable: a one-page “state of the numbers” snapshot, circulated by Friday.
Week 2: customers and market.
- [3-5] customer interviews or recent-deal reviews; [2-3] lost-deal reviews.
- Competitive and category scan: who is winning attention for our problem space, including in AI search results.
- Content and SEO/GEO audit: what ranks, what gets cited by AI assistants, what is dead weight.
- Deliverable: voice-of-customer notes and a competitive snapshot.
Week 3: synthesis.
- Funnel math assembled end to end: where volume drops, where quality drops, what a lead is worth.
- Positioning assessment: does the current message match what won the last ten deals?
- Deliverable: the audit document, with the two or three highest-leverage gaps named and sized. Presented live to leadership.
Phase 2: Decide (weeks 4-5)
Week 4: strategy draft.
- Proposed ICP focus, positioning, channel priorities, content/GEO plan, budget allocation, team and tooling implications.
- Explicit “what we will not do” list; a strategy without one is a wish list.
- Deliverable: strategy document, one page plus appendix.
Week 5: alignment and sign-off.
- Working session with leadership; disagreements resolved or parked explicitly.
- KPI tree agreed: the one number marketing owns, and the three or four beneath it.
- Deliverable: signed-off strategy and a quarterly metrics commitment.
Phase 3: Build (weeks 6-9)
- The operating cadence starts: weekly priorities, [biweekly] metric reviews.
- First experiments live by [week 7]: [e.g. positioning test on the homepage, one outbound-assist campaign, two GEO-targeted content pieces].
- System assembly, scoped to the strategy: analytics fixed to track the KPI tree, content pipeline running, [hiring or agency selection if in scope].
- AI leverage installed where it pays: research workflows, content production system, reporting automation. [This is where an AI-native operator visibly differs; see the operating system.]
- Deliverable: a live dashboard of the KPI tree and at least [2-3] experiments with written readouts.
Phase 4: Scale and decide (weeks 10-13)
- Double down on what showed signal; kill what did not, in writing.
- Playbooks documented so the team can run them: [briefs, channel runbooks, reporting cadence].
- Team plan forward: what to hire, what to contract, what the operator keeps owning.
- Deliverable: the renewal memo: results against the week-5 commitments, honest variance notes, and a proposed scope for the next quarter.
How to use this template as a company
- Hand it to candidates. Asking “mark up this 90-day plan for our situation” in the final interview round is more revealing than any portfolio review; pair it with the interview questions.
- Calibrate the gates to your sales cycle. With a 6-month enterprise cycle, demanding pipeline-to-close attribution by week 13 is math denial. Trend-level leading indicators are the honest ask.
- Inspect artifacts, not hours. The retainer buys judgement and a system (what retainers cost and include); the artifacts are how you verify the system is real.
- Use the final question as the bar: if the operator disappeared after week 13, would the strategy, dashboard, cadence and playbooks keep working? If yes, you bought an asset, not a dependency.
If you want this template adapted to your stage and motion, my DMs are open on LinkedIn.
FAQ
What should a fractional CMO accomplish in the first 90 days?
Four phases: a diagnosis of the funnel and positioning in weeks 1-3, a signed-off strategy by week 5, a working marketing system with live experiments through weeks 6-9, and scaled execution with a renewal decision basis by week 13. Each phase has concrete deliverables, not just activity.
What are typical fractional CMO deliverables in the first month?
A funnel audit with numbers, an ICP and positioning assessment, a channel and budget review, interviews with sales and customers, and a prioritized list of the two or three highest-leverage gaps. In writing, presented to leadership.
How quickly should a fractional CMO show results?
Diagnosis and strategy within a month; first revenue-relevant experiments live in month two; trend-level signal on pipeline metrics by the end of month three. Demanding closed revenue inside 90 days from a standing start usually means the sales cycle math was ignored.
How do I evaluate a fractional CMO at the end of 90 days?
Score against the plan’s gates: was the diagnosis specific and numerical, did leadership sign the strategy, are experiments live with honest readouts, and would your team keep the system if the operator left tomorrow. In my experience the last question is the strongest single indicator.