In January 2026, mobile gaming ad spend rose 42 percent year over year. Gaming downloads over the same month rose 2 percent, around 21 million new installs. Those are both Sensor Tower numbers, from the same report, covering the same month.

Spend grew roughly twenty times faster than the installs it bought.

That is not a bidding problem you can optimise your way out of. It is price discovery against a fixed pool of users, and every quarter you stay inside the same three self-serve auctions, you pay more for the same person.

The channel most teams skip is already most of the market

Here is the part that makes the whole thing awkward. Liftoff and Singular found that about half of casual game installs already come from ads displayed in hypercasual and puzzle genres. App to app is not an alternative to paid UA. It is already the bulk of paid UA.

The only real question is whether a network sits in the middle, taking margin and owning the attribution, or whether the relationship is direct.

App to app affiliation covers three distinct models, and they behave differently.

1. Reciprocal cross-promotion. Two apps with the same ideal user and no competitive conflict promote each other through in-app placements. At comparable audience sizes the media cost is close to zero.
2. Performance affiliation. A publisher app promotes yours and is paid per verified action. CPA by construction, which is why the incentives line up.
3. Owned-ecosystem cross-promotion. If you run more than one app, you already own an acquisition channel and are probably renting those users back from a network you paid to acquire them from in the first place.

What the price data says about staying put

Adjust’s Gaming App Insights Report 2026, published 26 March 2026, puts global gaming CPI at 0.56 dollars in 2025, up 30 percent year over year. Regionally: North America moved from 1.28 to 1.68 dollars, up 31 percent. Europe rose 47 percent, from 0.36 to 0.53. LATAM grew 40 percent to 0.14. By subgenre, slots reached 4.47 dollars and idle RPG 3.19.

The more revealing number in the same report is the paid-to-organic install ratio. The global median went from 2.07 to 3.33, a 61 percent jump in one year. Casino titles hit 11.05, up 223 percent.

Advertisers are buying more paid installs per organic install, and paying more for each one. AppsFlyer’s State of Gaming for Marketers 2026, built on 9,600 apps and 24.8 billion installs, adds the supply side of the squeeze: paid install share rose 10 percent year over year while ad impressions rose 20 percent. More ads competing for the same players.

Quality, measured properly, is where the direct channel wins

Install-level CPA will always make a direct partnership look expensive next to a network CPI. That comparison is the trap.

AppsFlyer’s owned media analysis found that conversion rates from owned media run close to double those of paid media, and the paying-user gap is starker: for gaming in Europe, 7.5 percent of owned media users became paying users against 2.2 percent from paid media. That data is from 2024, and I am flagging the year deliberately, because it is the cleanest owned versus paid quality contrast anyone has published and nothing newer has replaced it.

The lesson survives the date. Direct and owned traffic does not win on cost per install. It wins on what the installed user does next.

What to do next week

4. Audit your paid-to-organic ratio by country before you approve another budget increase. If your ratio is above the Adjust median for your genre and still climbing while total installs stay flat, incremental spend is buying users you were partly getting anyway.
5. Build a per-geo, per-genre CPA grid from your own CPI, using the Adjust benchmarks as the ceiling. A single blended global CPA offered to a partner is either uncompetitive in North America at 1.68 dollars or ruinous in LATAM at 0.14. The spread is 12x. One number cannot serve both.
6. Set up the partner cohort in your MMP before the first partnership install lands, and measure paying-user share, not installs. That split cannot be reconstructed retroactively, and installs are exactly the metric that hides the difference.
7. Start with two partners, not twenty. This channel scales through relationships, not through a budget slider, and the operational load is real.

The limitation

Cross-promotion moves users around. It does not create them.

The academic version is blunt. Haenlein, Libai and Muller, writing in the International Journal of Research in Marketing in 2023, frame cross-promotion as the industry’s answer to satiation: engagement decline that lowers user lifetime value to the point where acquisition from the large platforms stops making sense, “driving game publishers to cross-promote, that is, sell and swap users among themselves”. Useful, and genuinely valuable, but it is a reset mechanism, not net new audience. A portfolio-only strategy eventually compounds its own satiation.

The second limit is operational. Adjust found gaming apps consolidating to an average of 5.3 network partners in 2025, down from 6. Direct partnerships run against that trend, because each one is negotiated, integrated and reconciled by hand. This is a margin and quality play at the top of your funnel. It is not a volume replacement for the self-serve networks, and anyone selling it as one is overselling.

Why the model matters more than the channel

A partnership only aligns if the partner is paid for the outcome rather than for the impression. That is the whole mechanism. Under CPM or CPI, a partner’s incentive is volume, and volume is exactly what is already inflating in the auction.

For example, at SpinX the model is CPA with MMP-verified attribution, on inventory we own and operate plus direct publisher relationships. That structure exists because it is the only version of app to app that survives contact with a finance team: the advertiser pays per verified action, the risk of weak traffic sits with us, and the measurement lives with a third party neither side controls.

If your CPI keeps climbing while your install volume does not, it is worth pricing a direct, pay-per-result partnership against your current blended cost at spinx.io.