In-house is cheaper per ad and worse at absorbing spikes; an agency costs three to five times as much per ad and can double your output next week without you hiring. Compare the two on cost per published ad rather than per produced ad, and on how much your volume can move in a fortnight. Those two numbers decide it, and almost every other argument in this debate is about org charts.
The comparison people actually make is the invoice against a salary, which is the one comparison guaranteed to mislead. A retainer is fully loaded and visible. An in-house team's cost is hidden across four budgets and its ceiling is invisible until you hit it.
Cost per published ad, worked properly
Two inputs make the numbers honest. The first is loaded hourly cost, meaning salary plus employer overhead rather than the salary line alone, which for a mid-level editor in Western Europe lands around EUR 45 an hour. The second is keep rate, published divided by produced, because you pay for everything you produce and only benefit from what you publish.
In-house editing of a short-form ad from existing footage typically runs 30 to 90 minutes, so roughly EUR 25 to 70 of loaded time per produced ad. Agencies commonly charge EUR 100 to 400 per ad depending on scope, and whether they shot the footage is usually what decides where in that range you land.
Now assume 40 published ads a month, which is a realistic weekly test cadence for one active account.
| Line | In-house editor | Agency at EUR 150 per ad |
|---|---|---|
| Keep rate you see | 60% (you review everything produced) | 85% (they filter before delivery) |
| Ads you pay for | 67 | 47 |
| Production cost | 67h at EUR 45 = about EUR 3,015 | about EUR 7,050 |
| Briefing | 8 briefs, 4h, about EUR 180 | 8 briefs, 6h (longer, written for outsiders), about EUR 270 |
| Your review time | 5.6h, about EUR 252 | 3.9h, about EUR 176 |
| Launch and QA | 3h, about EUR 135 | 3h, about EUR 135 |
| Monthly total | about EUR 3,580 | about EUR 7,630 |
| Cost per published ad | about EUR 90 | about EUR 191 |
| Time to double output | 4 to 8 weeks (hire) | 1 to 2 weeks (invoice) |
Notice which lines move. In-house wins on production cost by a factor of two, and loses on briefing and on elasticity. Notice also the keep rate difference, which is the part most comparisons omit entirely: a good agency delivers pre-filtered work, so your visible keep rate is high because their internal one is invisible and priced in. Our video ad cost calculator will run this with your own hourly figure and your own keep rate, which is the only version of this table worth acting on.
The keep rate question, for both routes
Keep rate is where each route hides its real cost, and the honest version differs.
In-house. You see every output, so your keep rate is genuine and usually lower than people expect: 55 to 75 percent when there is a written brief, and down in the forties when brand and legal constraints are discovered during review rather than stated in the brief. The upside is that a visible keep rate is a fixable keep rate. You can trace a rejected ad to a missing brief field and fix the field.
Agency. Your visible keep rate is high, often 80 to 90 percent, because you are reviewing a shortlist. Their internal keep rate might be 50 percent and you are paying for it inside the per-ad rate. That is not a scandal, it is what you are buying, but it means "our agency's work almost always passes review" is not evidence that the agency is efficient.
The practical consequence: if you are in-house and your keep rate is under 50 percent, do not go shopping for an agency. Fix the brief first. Moving constraints from review into the brief typically takes keep rate from the forties into the seventies, which cuts your effective cost per published ad by around 40 percent without changing anything else. Keep rate is worth measuring for four weeks before you make a structural decision, because it will change which column of the table above looks better.
Volume elasticity is the agency's real advantage
Cost per ad is where agencies look weak. Elasticity is where they earn the premium, and it is the argument I find genuinely persuasive.
One in-house editor caps out around 40 to 60 produced ads a month before quality slides, and that ceiling does not move with budget. Adding capacity means recruiting, which is 4 to 8 weeks and a permanent cost. An agency absorbs a spike in a week or two against a purchase order.
Cases where the agency is clearly the right answer:
- A launch or seasonal peak where you need 150 ads in three weeks and 30 a month thereafter. Hiring for the peak leaves you overstaffed for ten months.
- A capability you need three times a year. Shoots with talent, licensed music, a specific animation style. Buying that in is correct.
- Markets you do not staff. Native-language creative for six markets is not something one person does well.
Cases where in-house is clearly right: a fixed weekly cadence, a category you understand better than any outside team, an offer that changes often enough that briefing an external team costs more than the editing, and any situation where the feedback loop from result to next brief needs to close inside a week. Our agency creative workflow page describes the process from the agency side, which is worth reading whichever chair you are in, because it tells you what you are actually buying.
The decay rates make the elasticity point sharper. Our 2026 ad fatigue benchmark puts CTR decline at 15 to 20 percent in a creative's first two weeks, with week three landing 45 to 70 percent below the launch baseline. A team that cannot flex output is a team whose creative is dead by the time the next batch is briefed.
The hybrid most teams actually end up with
Almost nobody runs pure in-house or pure agency after two years. The arrangement that keeps reappearing splits the work by decision type rather than by craft:
| Work | Who does it | Why |
|---|---|---|
| The offer and the angle | In-house | It depends on margin, roadmap and what the last test said |
| The brief | In-house | Whoever reads the results should write the next brief |
| Weekly variation volume | In-house or tooling | High frequency, low unit value, tight feedback loop |
| Shoots and net-new footage | Agency or production partner | Episodic, needs kit and crew |
| Peak and launch volume | Agency | Elasticity you cannot hire for |
| Native creative in new markets | Agency, then in-house once it works | Local judgement first, repetition later |
| The read of results | In-house | Never outsource the decision about what to do next |
The line that matters in that table is the last one. Outsource production if the arithmetic favours it, but the read of results, and therefore the next brief, belongs with whoever owns the account. Every dysfunctional agency relationship I have seen shares one feature: the agency owns the creative decision and the client owns the spend.
Where tooling changes the arithmetic
This is our product, so read it with that in mind. Generation from footage you already own moves the production line rather than removing it. Genyad charges 1 credit per variation, about EUR 1.52 on the EUR 99 Growth pack, with editing, re-exporting and uploading footage free, so the 67 produced ads in the table above cost roughly a hundred euros in credits instead of 67 hours. What does not change is the 5.6 hours of review, the briefing, and the launch time. Review becomes the whole cost, which is why keep rate stops being an accounting detail and starts being the number you manage.
Two honest limits. It only works if you have usable footage: no library, no output, and no tool substitutes for a shoot day. And it does not replace what an agency is best at, because there is no shoot, no talent, no licensed music, no AI avatars or synthetic presenters, no static banners, and no predicted performance score to tell you which output will win. Agencies also get something we do not offer on self-serve plans: client sub-accounts, consolidated invoicing and client-facing review seats are not there, so an agency using Genyad exports files and reviews them in whatever it already uses with clients. Our pricing page has the packs, including a free tier of 5 variations without a card, which is enough to test the arithmetic on your own footage before committing to a structure.
Frequently asked questions
Is in-house or agency cheaper for ad creative?
In-house is cheaper per ad, typically about EUR 90 per published ad versus around EUR 190 at a mid-market agency rate, using a loaded hourly cost of about EUR 45 and a 60 percent keep rate. The gap narrows if your keep rate is poor, because in-house pays for every rejected output while an agency's rejects are priced into their per-ad rate. Run it on your own numbers before deciding.
What is the biggest advantage of using an agency for creative?
Volume elasticity. An agency can double or triple your output in one to two weeks against a purchase order, while adding in-house capacity means a 4 to 8 week hire and a permanent cost. For launches, seasonal peaks and capabilities you need three times a year, that flexibility is worth the per-ad premium.
How does keep rate change the in-house versus agency comparison?
It changes it substantially, because you pay for produced ads and benefit only from published ones. An in-house keep rate in the forties can make in-house more expensive per published ad than a filtered agency delivery, while a keep rate in the seventies makes in-house roughly half the cost. Measure yours for four weeks before restructuring anything.
Can one person replace an agency for video ad creative?
For a fixed weekly test cadence built on footage you already own, often yes, at roughly six hours a week. For shoots, talent, licensed music and sudden volume spikes, no, and pretending otherwise is how in-house teams end up producing worse ads more slowly. Most teams settle on a hybrid where the offer, the brief and the read stay in-house and episodic production is bought in.