Hire a head of growth when your growth loop works and needs to work harder. Hire a fractional CMO when you cannot yet say which loop deserves the effort.
A head of growth is a permanent salaried operator who owns throughput. A fractional CMO is rented judgment that owns which machine you are building in the first place.
The founder who made me think hardest about this had a document open when we got on the call. The title said Head of Growth. Under it sat eleven bullets.
The first four were what you would expect: own the acquisition number, run the experiment calendar, report weekly on cost per customer. Bullet five said define our position in the category. Bullet nine said decide whether we move to usage-based pricing.
Nothing about the document was careless. She had written down what the company needed in the order it occurred to her, and somewhere around bullet five it quietly became two jobs stapled together.
Against every other way to buy growth leadership, this role gets one row in the wider comparison. It is the row founders get wrong most often.
Should I hire a head of growth or a fractional CMO?
A loop is the repeating path a stranger takes to becoming a paying customer, plus whatever makes each customer produce the next one.
A head of growth makes the loop you already have run faster. A fractional CMO decides whether that loop is the one worth running.
That makes it a question about location. A constraint sits inside that circle or above it, and the two places want different people in the chair.
Inside means the loop runs and does not run well enough. Checkout converts at 2% and ought to convert at 4%. Week-one onboarding leaks and nobody has instrumented where.
Those are throughput problems with a known shape. The fix is somebody who runs the same disciplined cycle every week for years and refuses to get bored by it.
Above means the loop itself is the open question. Whether this customer is worth building the company around. Whether the price can carry the cost of acquiring anybody.
Running a faster test cycle at questions like those gets you to a wrong answer sooner, with better dashboards attached to it.
Here is how the common mistake looks from inside the building. You hire a strong head of growth for a problem that lives above the loop, and because they are good they start fast.
Conversion improves, activation improves, and two quarters later revenue is about where it started, because every win landed on a funnel aimed at buyers who were never going to renew.
Nothing in the readout could have warned you. It measures the loop, and the loop was the thing in question.
Is the gap inside your funnel or above it?
The test for this takes twenty minutes. Open wherever your growth ideas accumulate and take the last thirty items.
Sort each one into two piles. The first holds what you could run tomorrow if you had the hands: fresh ad concepts, a shorter checkout, a win-back email. The second holds the items that are actually arguments about which funnel to run at all.
Read your own backlog. If every item is a test inside one funnel, you want a head of growth. If half the items are arguments about which funnel, the gap sits above the funnel.
Arguments are what goes stale in a backlog, because nobody present can close them. If the same question has survived three quarters of planning, count it three times.
The other complication is founder-shaped. Someone who has been holding the frame together settles those arguments in their own head on the drive home, so their second pile looks empty.
If that describes you, count your calendar instead. How many decisions last month could only have been made by you? That number is the job you are actually trying to fill.
If both piles come up thin and no channel repeats yet at a cost you can predict, it is early for either hire. The next dollar belongs in the founder's own hours.
Where does a head of growth's authority stop?
Reporting lines tell you more than titles ever will. A head of growth reports into marketing or product, a level below the executive who owns the plan.
So the scorecard arrives already written: signups, activation, cost per acquisition, retention at day thirty, each with a target beside it. They will move those numbers, and they should.
What they rarely do is take the position that pricing is why acquisition looks expensive. Price belongs to somebody else, who is sitting across the table.
A head of growth handed a strategy mandate will quietly do it badly rather than tell you it sits above their seat.
The result shows up in the artifact. You asked for a positioning refresh, and six weeks later you have a messaging document with three directions in it and no recommendation.
A recommendation is a decision, and the decision was never theirs. The work sits in a folder until somebody with the standing to choose opens it.
That ceiling is why sequencing sits above the seat. Holding an order of repairs while every channel owner argues for their own line is the spine of the operating system I run.
When is a head of growth the obvious hire?
Often, and that is worth saying plainly given who is writing. Three situations where the permanent seat wins.
The first is a loop that works and is starved of hands. The offer is settled, retention is verified, one channel repeats at a cost you can live with, and forty tests sit in the backlog with nobody to run them.
By year three that person has run four hundred experiments on your checkout and remembers how each went. You cannot buy that in slices.
The second is a company where the frame is already held: a founder who is genuinely a marketer, or an executive who already owns positioning and price. A second strategic layer just gives you two people debating direction.
The third is a deliberate bet on institutional memory. What a head of growth learns about your customers stays inside the building and shows up next year.
Rented judgment walks out when the contract closes. A good engagement leaves a great deal behind, and still never as much as the person who stayed.
The order that works when you need both
These two are complementary far more often than they are alternatives, and the strongest sequence runs the rented layer first.
When I have built a growth function from nothing, the opening weeks rarely went on running the loop. They went on deciding which loop deserved the investment and making the numbers honest enough to prove it.
The move straight after that was handing the loop to whoever would own it for the next three years, with the customer and the scorecard settled. That is how I sequence a team from zero.
There is a difference in kind here that no job description shows you. One seat you are trying to fill permanently. The other you are trying to vacate.
My engagements are supposed to end, and the measure of one is whether growth keeps running after I stop showing up. A head of growth is the opposite instrument, bought for year three.
That permanence question also decides the fractional versus full-time choice.
Write the job description last
Go back to the document. Mark each bullet I or A, for inside the loop or above it.
If the A bullets are a small minority, delete them and post the job today. If they are the reason you opened the document, you have been trying to hire your way out of a decision that whoever takes the job then inherits.
The founder with the eleven bullets did both, six months apart. She spent the first stretch closing the arguments in her second pile: who the product was for, what it should cost, which channels got funded.
Then she wrote the description again. Seven bullets, every one of them a number with a target beside it, and the person she hired walked into a funnel that had already been aimed.
If you are staring at a job description that has quietly become two jobs, the sorting exercise above is the first thing I would run with you. Let's talk.