The Fractional CMO's First 90 Days: What I Actually Do (Not What LinkedIn Says)

Week 1–2: Audit Everything. Trust Nothing.

The first thing I do when I start with a new client is assume that every metric in every dashboard is wrong until I've verified it. Not because people are lying — because tracking breaks quietly and nobody notices. This is the most common pattern I find across all company sizes.

The audit covers four areas:

• What's being measured — Pull up every marketing dashboard. Ask: where does this number come from? How is it defined? Is it the same definition used across all reports? You'll find metric definitions that differ between GA4, the CRM, and the BI tool. You'll find KPIs being tracked that nobody acts on. You'll find metrics that stopped being tracked 6 months ago and nobody noticed because the report was still generating (with zeros or stale data).
• What's broken — Run the tracking audit: UTM coverage (what percentage of inbound traffic has proper source/medium/campaign parameters?), conversion event firing (are all key events actually firing in GA4 and the pixel?), CRM data quality (duplicate records, stale lifecycle stages, missing fields, unmapped sources). In my experience, UTM coverage below 60% is common. Conversion events missing for some channels but not others: also common. Both of these mean your attribution model is fiction.
• What's the real CAC — Not the CAC the marketing team reports. The real CAC, calculated from finance's actual spend numbers and sales' actual closed-won data, with the right time lag applied. Marketing CAC and finance CAC often differ by 40–80% because marketing uses one set of cost inputs and finance uses another. Unified on one definition before either number means anything.
• What's the actual pipeline — Sit with sales leadership and understand: what's converting, what's not, what are the objections, where do deals stall, what does a good lead actually look like versus what marketing is optimizing for. The gap between marketing's definition of a qualified lead and sales' actual experience of that lead is always illuminating. Usually embarrassing for someone.

The Uncomfortable Conversations: What You Find and What to Do About It

The audit produces findings. Some of those findings are uncomfortable. This is where most fractional CMOs make their first mistake: they soften the findings, hedge the recommendations, and frame everything diplomatically to avoid upsetting the founder or CEO who hired them.

That's not what you were hired to do. You were hired to tell the truth.

The truths I've had to deliver in week 1–2:

• "Your CAC is $4,200, not $1,800. The $1,800 number excludes your two most expensive channels and doesn't account for the 3-month sales cycle lag."
• "The content marketing program has generated 2 attributed opportunities in 18 months and cost you approximately $180,000 including headcount. Here's why it's not working and what the path forward looks like."
• "Marketing is sending 800 MQLs per month to sales. Sales is working 60 of them. The other 740 are dead on arrival because the lead qualification criteria hasn't been reviewed since you were a Series A company and your ICP has changed."
• "Your marketing attribution model gives 100% credit to the last click. Your paid search is getting credit for leads that were already nurtured by 4 months of email sequences. You're about to cut your email program because it looks like it's not contributing to revenue. It's actually one of your most important channels."

These conversations are hard. But they're the reason you're there. The company doesn't need another person to validate the existing narrative. They need someone to tell them what's actually happening.

Week 3–4: Fix the Data Layer First

This is non-negotiable. Before any campaign changes, before any budget reallocation, before any new channel experiments: fix the measurement layer. You cannot make good decisions on bad data, and you will absolutely make bad decisions confidently if you don't fix it first.

What "fix the data layer" means in practice:

• UTM governance — Establish a UTM taxonomy document. Define the values allowed for source, medium, campaign, content, and term. Build it into a shared UTM builder tool (Google Sheets works fine). Require all traffic-driving activities to use it. Implement URL validation in the CRM form processing. This sounds basic because it is. It's also unfixed at most companies.
• Conversion event audit and repair — Test every conversion event in every channel. If a Facebook pixel conversion is double-counting (because GA4 and the pixel both fire on the same thank-you page), fix it. If a key event stopped firing after the CMS update, fix it. If the offline conversion import to Google Ads is stale because the data team changed the pipeline, fix it. Document what each event means and who owns it.
• CRM data baseline — Deduplicate records. Standardize lead source values (the CRM almost always has 40 variants of "organic" as a source — "Organic Search", "organic", "Organic search", "SEO", "Google Organic" — all meaning the same thing but splitting every report into noise). Map lifecycle stages to a clear definition that both marketing and sales agree on in writing. This is not glamorous work. Do it anyway.
• CAC definition alignment — Get finance, marketing, and the CEO in a room (or Zoom call). Agree on the exact cost inputs in the CAC calculation. Document it. Make it the official number. Stop using three different numbers in three different presentations.

The fix doesn't have to be perfect. It has to be defensible. A UTM coverage rate of 85% is good enough to make decisions. A conversion event that fires on 95% of actual conversions is good enough. You're establishing a measurement foundation, not solving for perfection.

Month 2: Build the Operating Model

With the data layer more reliable, the second month is about building the operating infrastructure that will let you run marketing like an actual function instead of a collection of disconnected activities.

The operating model covers:

• Rhythm — Weekly marketing-sales sync (not a one-way broadcast, an actual two-way exchange where sales tells marketing what's converting and marketing tells sales what's coming through the pipeline). Monthly full-funnel review with the CEO. Quarterly planning cycle that connects to the annual plan. These cadences don't exist at most growth-stage companies. The absence of a reliable operating rhythm is why priorities are constantly shifting and the marketing team is always reactive.
• Reporting structure — One weekly dashboard. Full-funnel, from impressions and traffic down to pipeline and revenue influence. Agreed metrics only — no vanity metrics in the weekly report. This dashboard is the single source of truth that all conversations start from. If the CEO wants to add a metric, great — it gets added to the dashboard, not discussed from someone's back-of-envelope calculation.
• Channel ownership — Who owns which channel? What's their goal metric? What's their budget? What's the review process if performance drops? Most marketing teams have vague ownership and no clear accountability structure. The operating model makes it explicit.
• Experiment backlog — A prioritized list of growth experiments with clear hypotheses, measurement plans, and success criteria. Nothing goes into this list without a hypothesis. Nothing launches without a success metric defined in advance. This is where the scientific rigor separates good marketing operations from random activity.

Month 3: First Experiments Ship

Only in month 3 do I recommend shipping net-new experiments. Before that, you're auditing, fixing, and building infrastructure. Any campaign or channel experiment before the measurement layer is clean is likely to produce misleading results that will drive bad future decisions.

Month 3 experiments should be scoped deliberately:

• Small, fast, measurable — Not a brand repositioning or a new channel launch. A landing page test. An email subject line A/B test. A bid strategy change in paid search. Something that can complete in 2–4 weeks and produce a clear result.
• Hypothesis-driven — Every experiment has a written hypothesis: "We believe that [changing X] will [produce result Y] because [rationale Z]. We'll know it worked if [metric] changes by [amount] in [timeframe]." If you can't write the hypothesis, you're not ready to run the experiment.
• Connected to the audit findings — The experiments should address the biggest leverage points identified in the audit, not the founder's gut feeling about what to try next. The audit usually reveals 2–3 high-confidence opportunities where the existing data suggests a clear improvement direction.

Month 3 is also when you should have enough data from the improved measurement layer to start making confident channel allocation decisions. You know your real CAC by channel. You know which leads are actually converting in sales. You know which content is driving pipeline versus just traffic. Now you can reallocate intelligently.

Why Most Fractional CMOs Fail

I've seen enough fractional CMOs fail (or more precisely, I've been brought into companies after a fractional CMO failed) to know the pattern. It's almost always one of three things:

• They executed before they diagnosed. They showed up with a playbook they used at the last company and started running plays without understanding the current company's specific situation. The playbook worked last time. It doesn't work here because the customer, the product, the team, and the measurement infrastructure are all different. Generic playbooks applied without diagnosis are expensive.
• They didn't fix the data layer. They made decisions based on the existing (wrong) numbers. Campaigns got cut that were actually working. Budgets went into channels that looked good but were measurement artifacts. The CEO lost confidence when results didn't match predictions because the predictions were based on bad data.
• They didn't manage the uncomfortable conversations. They found out that the marketing team was understaffed, underskilled, and working on the wrong things — and didn't say it clearly. Or they found out that sales was rejecting 90% of marketing leads and didn't surface that conflict to leadership. The fractional CMO's job includes being the person who says the true and inconvenient thing that internal people are too political to say. If you can't do that, you're not adding the value the engagement is priced at.

The 90-day framework above is not magic. It's structured common sense. Audit before acting. Fix measurement before making decisions. Build operating infrastructure before launching campaigns. Run experiments before making large bets. Most of the value a fractional CMO adds is in forcing this discipline on organizations that have been operating on intuition and hope.