What you get from a part-time CMO is defined by latency rather than hours: emergencies interrupt the day they happen, operating decisions come back the same day async, and strategy holds for a weekly session. The count of days matters less than that structure, and there are two situations where the model honestly fails.
Somewhere in every serious intro call, usually after the case studies and before the money, the founder gets to the real question. It arrives in different costumes. What does a typical week look like. How many clients do you carry. Who else gets your Tuesdays.
Underneath, it is always the same question: how much of you do we actually get?
The question deserves a better answer than the reassurance it usually gets, because the fear under it is legitimate. Founders who hired fractional leaders and hated it describe the same failure: attention spread so thin the leader never understood the business, everything slowed down, senior rates paid for tourist-level engagement.
So here is the mechanical answer I give on calls, including the parts that argue against the model.
Hours are the wrong unit
The category standard for engagements like mine runs one to three days a week. That number is true and nearly useless, because it describes the invoice rather than the experience.
The client never feels your hours. The client feels your latency.
Think about what you actually need from a marketing leader in a given week. A handful of decisions that only senior judgment can make. Fast unblocking of the team's work. Someone watching the numbers who knows which movements matter. One deep working session where strategy gets decided with the data on the table.
None of that requires forty hours of presence. All of it requires the right latency on each type of decision, which is a design problem, and designable.
Count it yourself at your next weekly: the number of decisions that genuinely required senior marketing judgment. Most companies in the $15M to $50M band produce three to six a week. The rest of a full-time leader's forty hours goes to presence, meetings attended in case something needs them, and the meetings expand to absorb the hours available.
How a week actually allocates
Tier one is the column founders worry about, so it gets the strongest guarantee: spend anomalies, tracking outages, and moving launch windows interrupt whatever I am doing, the day they happen. Alerts exist so that the account raises its own hand; nobody has to hope I happened to look.
The alert list itself is short and boring, which is why it works: daily spend outside an agreed band per channel, conversion events flatlining for a couple of hours, acquisition cost past threshold for two days running, any change to the tracking layer. Fewer than ten rules catch nearly everything that has ever actually mattered in an account I run.
Tier two is the daily texture of the engagement: approvals, in-band budget shifts, offer calls, vendor questions. These move async, in writing, the same working day. The team never waits until Thursday to ship Tuesday's work.
Tier three is deliberate slowness. New channel bets, pricing moves, hiring, strategy changes: these hold for the standing weekly session, with numbers on the table. Strategy decided reactively, between meetings, in a panic, is how good accounts get wrecked, and a part-time calendar is oddly protective here: it makes reactive strategy structurally hard.
What part-time buys that full-time cannot
There is one asset a fractional arrangement delivers that no salary can: current pattern exposure. I am inside multiple companies' numbers in any given season, watching the same platforms behave across different budgets and models. When your CPMs lurch, I usually know within days whether that is you or everyone.
The seat changes what panic looks like, too. When platform costs lurch, the first founder question is always the same: what did we break? From a multi-account view, whether the lurch is yours or the whole market's is visible within days, and half of marketing panic dissolves in that one distinction.
A full-time leader sees one company's data and last year's memories. Whether that trade favors you depends on your situation, and the full-time comparison is its own honest piece.
When is part-time not enough?
Launch mode is the first breaking point. A product launch, a rebrand, a peak-season campaign week: these need senior eyes daily, sometimes hourly, for a bounded stretch. The honest fix is naming that period in advance and planning intensity for it, and an engagement that pretends every week is the same week will fail exactly there.
This is a planning artifact rather than a heroic exception. At FX Replay the peak season ran as exactly that kind of scheduled intensity: a six-week runway agreed months out, with daily senior attention on the calendar instead of improvised at midnight.
The second breaking point is structural: nobody inside who can decide between sessions. That one is a gate I check before taking the engagement, because no latency design survives a company where every choice queues for one calendar.
What this costs you to make work
The model asks three things of the client, and they are cheap: an inside owner for day-to-day calls, real access to the data and accounts, and protected attendance at the weekly. In exchange the engagement runs on instruments and structure rather than on presence and hope.
Protected means protected. The weekly holds in bad weeks, especially in bad weeks, because skipping the review in a bad month is how bad quarters get assembled.
Presence is a poor substitute for instruments anyway. Full-time leaders also miss the silent failures when nothing is wired to ring. The difference is that a part-time engagement cannot pretend otherwise, so the wiring gets built first, and the wiring is what you keep after the engagement ends.
The latency table above is the actual offer, more than any hour count is, and the economics are published separately. If you want the table run against your company's real weeks, let's talk.
