A founder past product-market fit asked me a clean question. With budget for exactly one growth hire this year, does the money go to an agency or to a fractional operator? The honest answer starts by admitting that both can fail, and that they fail in opposite directions.
You have proven the product. Customers are buying, retention holds, and the machine that reliably finds the next thousand of them is the thing you do not have yet.
So you look outside, and the market hands you two shapes of help. An agency, a team that runs channels for a roster of clients. And a fractional operator, a senior marketer who owns your growth for a slice of their week.
They tend to cost about the same across a year. You can afford one. A wrong pick does not merely waste the budget, it burns the window when momentum was cheapest to compound.
I have stood on every side of this. I have been the fractional operator a founder bet on. As a CMO I vetted, hired, and managed agencies, and let go of the ones that were not working. Early in my career I worked inside one.
Here is how I reason it now: choose by how each one fails, because each one does.
What an agency is built to do well
An agency is a specialist at production and channel depth. It runs the same channel across dozens of accounts, so it has already met the failure modes, the account structures, the creative formats that are working this quarter.
It arrives with a bench: a media buyer, a designer, an analyst, an account lead, all available the week you sign. Hand that same scope to one in-house generalist and you wait months, then get a single person's range where an agency gives you five people's.
When the work is production against a clear brief, an agency is often the correct answer and the cheaper one. What follows is simply the one thing that model leaves out.
Agencies execute inside the frame they are handed. The expensive mistakes usually live in the frame itself.
Where an agency breaks
The break is structural, and it has nothing to do with talent. If the brief says scale paid social, the agency will scale paid social with real craft, and none of that craft asks whether paid social is where this business should be spending at all.
The frame, meaning the choice of channel, the offer, the audience, the order in which you attack them, is the layer with the largest dollar consequences. It is also the layer sitting outside the engagement.
One client I later worked with had been handing an agency roughly $15,000 a month while a mis-wired tracking setup was feeding back under five subscribers' worth of trustworthy signal, a story I walk through in full in the piece on broken attribution.
The agency was competent. Nobody had asked it to check whether the numbers it was optimizing toward were even real, because that question lived in the frame, and the frame was somebody else's job.
That is the agency failure mode in one line. Beautiful execution aimed at the wrong target.
The better the agency, the more efficiently it runs in the wrong direction, and the longer the dashboards stay busy enough to keep you from noticing that the direction was the problem all along.
Read the incentive before the case studies
There is a second structural pull worth naming, and it comes down to arithmetic more than character. Many agencies are paid a retainer plus a percentage of the media they manage.
The ANA's 2022 agency compensation study found 19% of marketers still paying commission on media planning or buying, nearly three times the rate for agency services overall. That structure quietly rewards more spend and a longer engagement.
Plenty of good agencies resist it, and many do. But the incentive is always in the room, and it never points toward the sentence you sometimes most need to hear.
An agency paid on a percentage of spend can be excellent and still never tell you to spend less. Read the incentive before you read the case studies.
An operator carries a different pull. Paid for judgment rather than a cut of the budget, and often holding some outcome or equity stake, they are rewarded when the economics work and when they eventually make themselves less necessary.
Operators are no saints either. The point is that money reaches the two roles differently, so they lean different ways, and you want to know the lean before you sign anything.
What an operator is built to do, and where they break
A fractional operator owns the growth profit and loss. They set the frame: which channels, which offer, in which order, against which unit economics. They sequence the work so the company is not pouring money into a channel that has not earned it yet.
This is the judgment layer that sits above execution, the part that decides what an agency should be briefed to do in the first place.
Spencer Stuart's 2026 CMO tenure study describes the top marketing job widening into what it calls a CMO-plus role, with titles shifting toward commercial and revenue ownership. I have written about what that engagement looks like week to week in the engagement guide.
Their failure mode is the agency's in a mirror. An operator can strategize with no production muscle behind them.
One senior person, however sharp, cannot personally buy the media, cut the creative, build the lifecycle flows, and stand up the analytics. Give an operator a mandate that needs ten pairs of hands and you get a sharp plan sitting on top of starved execution, decks and frameworks with almost nothing shipping underneath.
Strategy with no production costs you as much as production with no strategy. It just fails more quietly.
When I ran growth as a CMO across several business units, a large share of the job was directing agencies rather than replacing them: briefing them, holding them to the frame, and cutting the ones executing cleanly against the wrong target.
Pick by failure mode
With an agency you risk a strategy that drifts; with an operator, production that starves. Whichever risk your stage can absorb is the one to take.
Here is the actual call. If your frame is already right and what you lack is hands and channel depth, the agency is the efficient answer, and you can absorb the risk of drifting strategy because you are the one holding the strategy.
If your frame is unclear, or your channels are plateauing and you cannot yet tell whether the problem is the ad or the offer or the sequence, the operator is the answer, and you can absorb slower production because the expensive question right now is direction rather than volume.
The end state most companies land on
In practice the mature answer is usually both, in a specific order. An operator holds the frame and the number, and directs one or more agencies as the production bench beneath that frame.
The operator decides what to build and briefs it, the agency builds it at a depth no single hire could reach, and the operator holds them to the target and kills the work that wanders off it.
If you can only fund one this year, buy the layer you are missing, and buy it knowing the other layer is where you grow into next. This is the same logic behind the growth operating system: settle the frame first, then resource the execution against it.
The founder who asked me the clean question had a frame problem rather than a hands problem. Their channels had gone flat and nobody in the building could say why.
More execution against a target they could not yet trust would only have burned money faster. What was missing was an owner for the target itself.
A year on, with the frame settled and the economics legible, the agency conversation became the easy one it should have been from the start.
If you are staring at this choice with budget for one, the fastest way through is to name your failure mode first, then buy against it. That is the diagnosis I run with founders past product-market fit. Let's talk.