Direct answer: should mobile games use a UA agency or build in-house? Use an agency below roughly $300,000 a month in media spend, run the blended model between $300,000 and $500,000 a month, and build an in-house core above $500,000 a month. Every figure on this page is monthly media spend in US dollars, because that is the axis the arithmetic works on and mixing it with annual budget is where most published advice contradicts itself. The mechanism: agency fees scale linearly with spend, an internal team is a step function of fixed cost, and the two cross in a band rather than at a point. If you want that band computed against your actual P&L, that is the kind of question a gaming growth advisor settles in a session.
The UA agency vs in-house mobile games debate gets argued on the wrong axis. Studios frame it as a control question, when it is a fixed-versus-variable cost question with a computable crossover. In my experience across Gameloft, SFR, and Impulse Media Hub, the studios that got this wrong did not pick the wrong model. They picked the right model at the wrong spend level and discovered it nine months and a million euros later.
Key Takeaways
- One axis, one currency. Monthly media spend in US dollars. An annual budget of $5M is $417,000 a month, which sits inside the blended band. Sources quoting a monthly threshold and an annual threshold are usually stating the same crossover twice
- Under $10,000 a month in media spend, no specialist agency will take the account. A freelance user acquisition manager at $800 to $2,000 a month is the realistic option
- Fees are 10% to 20% of media spend for specialist user acquisition agencies. Zoomd, a full-service vendor, publishes 15% to 25%. Single-channel retainers run $3,000 to $8,000 a month on top
- The crossover is a band rather than a line. The fee overtakes a full in-house function somewhere between $420,000 and $830,000 a month in spend, depending entirely on the percentage you negotiated
- In-house fixed cost is above $1M a year for the full six-role function, per Zoomd, plus a 9 to 12 month ramp. The blended two-head core runs $250,000 to $350,000
- Creative volume moved the goalposts. Top gaming advertisers ship 2,400 to 2,600 creative variations a quarter, up 25 to 30% year over year. Two internal designers cannot hold that cadence
What an in-house UA team actually costs in year one
Zoomd, which sells into this decision and therefore has an interest in the number being large, puts it plainly: “Team salaries, benefits, technology, and overhead can exceed $1M in fixed costs”, and “hiring, onboarding, and building operational maturity typically takes 9 to 12 months.”
The figure only makes sense once you see the team it describes. Zoomd’s in-house structure is six roles: a user acquisition manager, a paid social specialist, a programmatic specialist, a creative manager, a data analyst, and a creator campaign manager. Six heads at $1M is roughly $165,000 a head fully loaded, which is a credible European or US number for senior performance marketing roles once employer contributions, tooling seats and recruiting are counted.
That check matters, because the version of this page I published in July claimed the same $1M for a team of three or four. That works out at $250,000 to $333,000 a head, which no studio pays for a media buyer, and it sat in the same article as a blended two-head team at $250,000 to $350,000, or $125,000 to $175,000 a head. The same user acquisition lead cost twice as much in one paragraph as in another. The fix is to name the headcount every time a fixed cost appears, which is what the table below does.
The line item nobody models is the ramp. If you are spending $500,000 a month during those 9 to 12 months and running 20% below the efficiency your agency was delivering, the invisible cost of learning is $100,000 a month, so $900,000 to $1.2M across the ramp, on top of the fixed cost. The 20% is my assumption rather than a measurement, and it is the single input most worth arguing about before you commit. Nobody publishes a measured ramp penalty, because the studios that lived it were not running a control group.
The upside is real and compounding: full strategic control, direct platform relationships, and institutional knowledge that stays in the building. It has to be paid for at a scale that makes the payment rational. The same fixed-versus-variable logic applies to a single senior hire, which I break down in consultant versus in-house hire.
What a UA agency costs in 2026
| What you are buying | 2026 range, USD | Notes |
|---|---|---|
| Management fee, specialist UA agency | 10% to 20% of media spend | The segment that runs mobile game campaigns |
| Management fee, full-service agency | 15% to 25% of media spend | Zoomd’s published range, broader service scope |
| Single-channel retainer | $3,000 to $8,000 / month | On top of media, one channel |
| Minimum monthly media spend accepted | $10,000 to $30,000 | Below this, agencies decline the account |
| Freelance UA manager | $800 to $2,000 / month | 1.5 to 4 days of work. Not an agency |
Two things in that table do most of the work.
The first is the floor. Below roughly $10,000 a month in media spend, the question is not which agency, it is whether anyone will pick up. Fee percentages do not cover an account team at that volume, so specialist shops decline. Studios in that segment are choosing between a freelance user acquisition manager and doing it themselves, and pretending otherwise wastes a quarter on pitches that end in polite no-thank-yous.
The second is that the fee is the honest comparison number. On a $3.6M annual media budget, a 15% fee is $540,000 a year. That is what you weigh against a loaded internal team, never against one manager’s salary. Getting this comparison wrong by a factor of four is the most common error I see in studio board decks.
A previous version of this page also quoted a “mobile-specialist floor of around $25,000 a month”. That figure was welded together from two unrelated numbers on an aggregator page: the bottom of an enterprise retainer tier, and one named agency’s client minimum. It contradicted two other rows in its own table. It is gone.
What you buy for the fee is not only labour. It is enterprise demand-side platform access, established creator networks, platform beta access, and pattern recognition across concurrent accounts. An agency running twenty active gaming clients sees a creative angle decay on TikTok before your internal team would, because it is watching the same decay in twenty places. Our TikTok UA playbook shows how channel-specific that gets.
The spend thresholds that decide it
| Monthly media spend | Annual media spend | Recommended model | Why |
|---|---|---|---|
| Under $10,000 | Under $120,000 | Freelance specialist | Below the floor at which agencies accept accounts |
| $10,000 to $300,000 | $120,000 to $3.6M | Agency plus one internal owner | Fee funds well under two heads |
| $300,000 to $500,000 | $3.6M to $6M | Blended, split by function | Fee funds a small core but not a full function |
| Above $500,000 | Above $6M | In-house core plus agency creative | Fee approaches or exceeds a full team’s cost |
Every row multiplies by twelve. That sounds like a low bar until you check the version of this page I published in July, where the last row read “above $500,000 a month” against “$5M to $10M a year”. Five hundred thousand a month is $6M a year, so the row understated its own floor and overlapped the row above it.
Run the crossover yourself
The thresholds above are defaults. Your crossover depends on one number I do not have, which is your negotiated fee percentage.
Set the annual fee equal to the $1M fixed cost of the full six-role function and solve for spend:
- At a 20% fee, the fee reaches $1M at $5M a year, so $417,000 a month
- At 15%, it reaches $1M at $6.67M a year, so $556,000 a month
- At 10%, it reaches $1M at $10M a year, so $833,000 a month
The crossover therefore spans roughly $420,000 to $830,000 a month. Anyone publishing a single crossover point is publishing one fee assumption and not telling you which. This is also why “build in-house above $5M a year” and “run blended between $300,000 and $500,000 a month” are not two rules. They are the same crossover stated on two axes, with $5M a year landing at $417,000 a month, inside the blended band.
Two caveats. These thresholds assume sustained spend rather than a launch spike, and three good months do not justify permanent headcount. They also assume you can hire. A user acquisition lead who can own creative strategy, incrementality testing and attribution hygiene is scarce, and an underpowered internal team loses to a competent agency at any budget.
Not sure which side of the band you sit on? Book a strategy call and we will run the crossover against your real spend curve and your actual fee.
Why creative volume changed the equation
The 2026 shift is that creative production became the scarce resource, and that quietly moved the crossover.
AppsFlyer reports that top gaming advertisers now produce between 2,400 and 2,600 creative variations per quarter, up 25 to 30% year over year. Paid install share rose 10% year over year across iOS and Android while ad impressions increased 20%, so the same players cost more to reach. Global gaming app user acquisition spend reached $25B in 2025, growing 3.8% year over year, with nearly half flowing into the US even as budgets there declined 5%. China-headquartered publishers captured 35% of global gaming user acquisition spend outside China, and their share grew 22% year over year. Competition intensified faster than budgets.
Hold the unit steady when you compare yourself against that. AppsFlyer counts variations, meaning individual renders rather than distinct creative ideas. At 2,400 to 2,600 a quarter, that is roughly 800 to 870 variations a month.
A small studio’s creative wave produces 6 to 10 variations from 2 or 3 concepts, at roughly $1,200 to $3,000 per wave. Run two waves a month and you ship 12 to 20 variations from 4 to 6 concepts. Against 800 to 870, that is a gap of 40 to 70 times in variation count.
The gap in genuinely distinct creative ideas is smaller and unmeasured, because no source publishes concept counts for top advertisers. The July version of this page converted AppsFlyer’s variations into “roughly 800 concepts a month” and compared them against “15 to 30 new concepts a month” for a studio. That comparison inflated the gap by an order of magnitude by changing the unit mid-sentence, and it is the exact failure mode our CPI benchmarks piece is about: published figures for the same market are routinely incomparable because nobody restates what was measured.
The counterweight is insight quality. Agency creative drifts generic when the team does not live inside your game’s economy, meta or community. That tension is what the blended model resolves, and it is why the creative strategy brief stays in-house even when the rendering does not.
The blended model
The blended UA model splits ownership by function rather than by channel. It is the default between $300,000 and $500,000 a month in media spend.
| Function | Owner | Rationale |
|---|---|---|
| UA strategy and budget allocation | In-house | Tied to LTV, roadmap, and P&L |
| Measurement, attribution, MMP | In-house | Data ownership is non-negotiable |
| Creative brief and hypothesis | In-house | Requires game and player knowledge |
| Creative production volume | Agency | Scale economics favour outside |
| Channel execution and bid ops | Agency | Cross-account pattern recognition |
| Vendor and network relationships | Shared | Studio signs, agency operates |
The rule I apply: outsource execution, never judgement. The moment an agency decides your budget split across channels, you have outsourced the one function that determines whether the operation is profitable.
Practically, blended means an internal user acquisition lead plus a data analyst, so two heads rather than six, supervising an agency retainer. Fixed cost lands at $250,000 to $350,000 a year, which is $125,000 to $175,000 a head loaded, in the same band as the per-head figure implied by Zoomd’s six-role structure. You keep the strategic layer and the data, and you get a clean migration path: as spend grows, pull functions in-house one at a time rather than attempting a big-bang transition during a live campaign.
How to run the decision in four questions
- Is your spend sustained above $300,000 a month? If not, the in-house arithmetic does not close. Revisit in two quarters
- What does your fee percentage make the crossover? Multiply annual spend by your negotiated rate, divide by your loaded cost per head. Under two heads, no internal team exists at that fee
- Can you fund a 9 to 12 month ramp? Model the performance gap explicitly. If your runway cannot absorb it, stay with an agency
- Who owns the data? Whatever you choose, the attribution contract, the raw event data and the creative library must be in your name. This is the term studios forget until the relationship ends
The last point deserves emphasis. Ask any prospective partner for named gaming references at your spend tier and genre, and ask what happens to your creative library and account access on day one after termination. A good partner answers both without hesitation.
If the question sitting underneath all of this is what senior outside help costs before you commit to either structure, game consulting pricing covers advisory fees in euros with the day counts attached. Advisory does not replace execution, and at the low end of the table above it is often the only outside help available.
What I could not verify
Agency fee structures are not published as a dataset. No primary source I found surveys mobile user acquisition agency pricing. The 10% to 20% specialist band and the $3,000 to $8,000 single-channel retainer are the ranges I see in proposals and reconcile against the reference set this blog maintains. Zoomd’s 15% to 25% is a published figure from a vendor with an interest in the comparison, which is why it is labelled as theirs.
The ramp penalty. The 20% efficiency gap during the 9 to 12 month ramp is my planning assumption. Zoomd publishes the duration, nobody publishes the magnitude.
Retainer tiers by agency size. An earlier version of this page carried boutique, mid-market and enterprise retainer bands. They came from an aggregator page that recycles figures from other aggregator pages, and I could not trace any of them to a source that measured anything. Removed rather than re-sourced.
Concept counts at the top of the market. AppsFlyer publishes variations. The number of distinct creative concepts behind those variations is not published by anyone, which means the honest statement of the creative gap is a range in variations and a shrug on concepts.
Conclusion
The UA agency vs in-house decision has a defensible answer at every spend level, on one axis. Below $300,000 a month in media spend, use an agency and keep one internal owner. Between $300,000 and $500,000 a month, run blended. Above $500,000 a month, build an in-house core and keep an agency for creative volume, after checking where your own fee percentage puts the crossover inside the $420,000 to $830,000 band.
What breaks studios is running a model 18 months past the point where the arithmetic flipped, because nobody re-ran it.
Deciding how to structure your UA operation? Book a strategy call to model agency, in-house and blended against your actual spend, or explore our growth advisory services.