A failed marketing leadership hire costs about a year of loaded compensation, plus a search fee near a third of first-year pay, plus the item that appears on no budget: the stalled year. Every path to marketing leadership has a failure cost like that, nobody selling any path publishes theirs, and pricing them side by side changes which option looks expensive.
Walk the standard version first, because it has a schedule. Month three: optimism, the strategy deck lands well, everyone repeats its vocabulary. Month six: the numbers have not moved and the explanations begin, reasonable ones, delivered confidently.
Month nine: the quiet conversations start, the ones held about a person rather than with them. Month twelve: separation, and a search that starts from zero, in a market that now hears your company churned its last marketing leader.
The pattern is documented enough to be a cliche: in the early-stage version, the marketing VP departs inside a year, by resignation or removal. What almost never gets written down is the invoice.
Part of why it hides: the lines sit in four different ledgers. Payroll holds the comp, the search firm holds its fee, the separation agreement holds its own number, and the biggest line, the stalled year, sits in no ledger at all.
What does the failed hire actually cost?
Four lines. The seat: a year of loaded compensation at the executive tier, a range the full-time comparison prices honestly. The search: retained fees run about a third of first-year compensation, and a restarted search pays it twice. The exit: severance and transition, varying by contract and jurisdiction.
Then the fourth line, the one that dwarfs the ledger while appearing nowhere in accounting: the stalled year. Channels drifted. The team shipped without direction, then shipped less. Competitors ran their playbooks against a company whose growth function was busy having a leadership crisis.
The cash lines are recoverable in the next budget cycle. The year is not.
The other three paths fail too
Honesty requires pricing the alternatives the same way, including the one I sell. So, disclosure before the table: I sell the fourth row. The defense against my bias is that every range below is either linked to a public source or is arithmetic you can redo yourself, and the board framing at the end works whichever row you choose.
The agency year fails expensively when detection is slow. Twelve months of retainer is the visible line; the directed ad spend behind it is usually larger, and the retainer's own decomposition explains why performance problems surface late by default. With a real scoreboard, the same path fails cheaply at month three.
That detection speed is buyable separately from the agency itself. The instrumentation that grades an agency is the same scoreboard that grades anyone, it is cheap, and it converts the agency row from the slowest-detecting path into one of the fastest.
The founder-does-it path books no invoice at all, which is how it becomes the most expensive row on the ledger. The cost is the founder's own hours at the company's true top rate, spent on work a specialist does better, an argument made properly elsewhere. Its time-to-know is the slowest, because there is no one to compare against and no one positioned to call it.
This row also compounds in a way the others do not. Every quarter the founder runs marketing is a quarter nobody builds the evidence that would justify the right hire, so the path quietly extends itself.
And the fractional path fails too: wrong operator, wrong fit, wrong moment. What differs is the shape of the failure. Month-to-month terms mean the wrongness costs one to three months of retainer before the exit ramp, and a competent engagement states a prediction it can be graded against inside the first quarter.
Why time-to-know governs the ledger
Failure cost is a rate multiplied by a duration. Founders negotiate the rate hard and let the duration float, and the duration is where the money goes: by the time doubt becomes speakable in the full-time arc, month six or so, most of the year's cost is already committed.
So whatever path you choose, the highest-leverage clause you can negotiate is the one that shortens time-to-know: a stated prediction, a defined checkpoint, and instruments that can actually grade it. Those clauses cost nothing and reprice every row of the ledger in your favor.
In practice they look like: a 90-day checkpoint written into the agency contract with named metrics. A first-quarter prediction agreed with the full-time candidate before the offer goes out. A fractional engagement that opens with an audit whose findings become the prediction. Same clause, three costumes.
The prevention layer sits one step earlier and is also free: most mishires are visible in the interview process, if the exam tests diagnosis instead of performance.
The board conversation
Boards are professional risk-pricers. What makes them nervous is a single option presented with no failure case, because everyone at the table has watched that movie.
So bring the ledger instead of the option. All four paths, priced by cost-when-wrong, time-to-know, and exit friction. Name your choice, state the prediction it will be graded on, and put the checkpoint date in the deck. That framing gets approved, and more usefully, it gets defended six months later, because the checkpoint was part of the original ask.
If the path you are pricing is the fractional row, the economics are published, and so are the situations where I am the wrong answer. Let's talk.
