In-House vs Agency vs Hybrid: The Real Cost Math
The real cost of in-house versus agency versus hybrid, past the retainer: salaries, ramp, tools, and where the crossover actually sits by spend.
By The Spend Report Editorial Team. Published June 26, 2026. · 7 min read
On this page
When a founder tells you the agency is too expensive at $8K a month and they're going to "just hire someone," they're almost always comparing the wrong two numbers. The $8K retainer is a real, visible line item. The in-house alternative they're picturing is a salary they've already discounted by about 40 percent in their head. The actual decision is not retainer versus salary. It is fully loaded cost versus fully loaded cost, and the loaded numbers behave very differently as your spend grows.
Get this comparison wrong and you make one of two predictable mistakes: you keep paying a percentage-of-spend retainer long past the point where it stops making sense, or you yank everything in-house at a spend level that cannot support the headcount, and your performance quietly degrades for two quarters while you figure that out.
The number you're actually paying
Start with the agency side, because it's the one people misread first. A retainer is rarely the whole bill. If the agency prices on a percentage of spend, your cost scales with the thing you're trying to grow, which is fine at $80K a month and painful at $400K. Flat retainers invert that problem: cheap at high spend, expensive at low. Either way there are usually creative fees, a tooling pass-through, and the soft cost of the hours your team spends managing the relationship. The way agencies actually price determines which of these costs dominates, so read the structure before you read the number.
Now the in-house side, which is where the discounting happens. A "$95K media buyer" does not cost you $95K. Payroll tax, benefits, equipment, software seats, and the fraction of a manager's week that person consumes push the fully loaded figure to roughly 1.3 to 1.5x base. That same buyer is also not productive on day one. Ramp on a new channel owner runs one to three months before they're steering spend as well as the person they replaced, and during that window you're paying full freight for partial output.
| Cost component | In-house | Agency | Hybrid |
|---|---|---|---|
| Base salary or retainer | Visible | Visible | Both, smaller each |
| Payroll tax + benefits (~30-50% of base) | You pay | Baked in | On your headcount only |
| Ramp time (1-3 mo at partial output) | You eat it | Faster start | Split |
| Tool + data stack | You buy all seats | Often included | Shared |
| Management + QA time | Your hours | Their hours | Mostly theirs |
| Coverage when someone quits | Single point of failure | Bench absorbs it | Bench absorbs it |
| Knowledge that walks out the door | High risk | Process retains it | Lower risk |
The two columns people forget are the bottom two. A single in-house owner is a single point of failure: when they leave, your account knowledge, your test history, and your channel relationships leave with them, and you're back to ramp. An agency's bench absorbs that. That continuity is worth real money and almost never shows up in the side-by-side a founder sketches on a napkin.
Where the curves cross
Plot fully loaded cost against monthly ad spend and the three models stop looking like competitors and start looking like phases. At low spend, in-house is the most expensive option per dollar managed, because one salary is a large fixed cost spread over a small budget. As spend climbs, that fixed cost gets diluted and in-house gets cheaper per dollar. A percentage retainer does the opposite: it stays cheap while spend is small and climbs in a straight line forever.
The agency and in-house curves cross somewhere in the middle band, often in the $120K to $180K monthly spend range for a brand running two or three paid channels. The exact crossover depends on your channel count, your retainer structure, and what salaries cost in your market, so treat the numbers above as the shape of the relationship rather than your specific answer. Run your own blended efficiency through the MER calculator before you trust any threshold, including this one.
The useful read isn't the single crossover point. It's that the hybrid line spends most of the middle band underneath both pure options. That's not an accident.
Why hybrid usually wins the middle
Hybrid means you own the parts that compound internally and you rent the parts that need scale, specialization, or coverage. In practice that's a small in-house core, often one growth lead plus one channel owner, paired with an agency or set of specialists for the channels where outside reps, fresh creative volume, or bench depth matter.
You get three things from that split. You keep the strategic context inside the building, where it informs product, merchandising, and retention instead of living in someone else's Slack. You get a bench for free, so a single departure doesn't reset you to zero. And you cap your fixed payroll while still buying specialist depth on the channels that move fastest. The cost curve sits low in the middle because you're paying for fixed capacity only where it pays off and variable capacity everywhere else.
The tradeoff is real and worth naming: hybrid has the highest coordination cost of the three. Two parties touching the same account need clear ownership lines, or you get the worst of both, with the agency waiting on your lead and your lead assuming the agency has it. If you can't staff a competent internal owner to run the relationship, hybrid degrades into an expensive agency engagement with extra meetings. The model only works when your in-house person is senior enough to direct the outside team, not just relay tickets to it.
Running your own math
Before you move anything, do four things in order.
Audit your true loaded cost
Retainer plus media, or salary times ~1.4 plus tools and ramp
Find your crossover band
Where does in-house stop being more expensive per dollar?
Decide what compounds internally
Own strategy and data, rent scale and specialization
Move one channel at a time, hold each to a CAC target
First, audit what you actually pay today, loaded. For the agency, that's retainer plus pass-throughs plus your team's management hours. For in-house, it's base times roughly 1.4 plus tools plus the ramp you'll eat. Put both on the same basis or the comparison is fiction.
Second, find your band. If your spend sits well below the crossover, a pure agency or a lean hybrid is almost always cheaper per dollar managed, and hiring a full channel owner is premature. If you're well above it and running stable channels, in-house starts to pay for itself, provided you can solve coverage. Most brands doing $80K to $250K a month land in the hybrid sweet spot, which is exactly why hybrid is so common at that size.
Third, decide what compounds. Strategy, your data layer, and customer knowledge should live inside the building regardless of model. Raw execution hours on a fast-moving channel are the most rentable thing you own. When you do hand work off, knowing how to read an agency performance report is the difference between managing the relationship and being managed by it.
Fourth, move one channel at a time and hold each to a CAC target before you add the next. The hidden costs of switching agencies and the cost of a botched in-house transition are the same cost: lost learning, relearned accounts, and a quarter of degraded performance while the new owner ramps. You avoid both by transitioning narrowly, not all at once. If you're standing up the internal core for the first time, the sequence for building a growth team matters more than the model label you put on it.
The honest bottom line
There is no structurally cheapest option, only a cheapest option for your spend, your channel mix, and your tolerance for single points of failure. Pure agency wins when spend is low or volatile and you can't justify fixed headcount. Pure in-house wins when spend is high and stable and you can staff for coverage. Hybrid wins the long middle most growing DTC brands actually live in, because it dilutes fixed cost, keeps the compounding work inside, and buys you a bench, at the price of coordination you have to be senior enough to run.
Compare loaded to loaded. Find your crossover band with your real numbers. Move one channel at a time. The structure that looks expensive on the retainer line is often the cheapest one you can actually run.