Spot the Triple — How a Bird-Sorting Puzzle Buys Its Players on Ad Networks

BirdFly: Sort Puzzle launched on September 29 and opened 367 ad plans in six weeks, with the monitored budget running through AdMob, Unity, AppLovin, YouTube, Mintegral and two more networks — the monetisation side of the market, where casual ad-monetised titles do most of their buying. No Meta plan appears in the monitored window. The plan count is unremarkable; what stands out is the channel mix and a playable-heavy creative line.

Where the budget goes: the monetisation side of the market

BirdFly: Sort Puzzle is a sorting game from Casual Games For Fun, released on September 29, 2026. In the thirty-one days to November 9 it ran 367 ad plans against 308 distinct creatives — almost one plan per creative — the profile of an account that spins up a new plan for each new asset instead of scaling a handful of winners.

The channel mix is where this account breaks from its peers. AdMob covers 42.78% of plans, Unity 28.34%, AppLovin 16.62%, YouTube 15.80%, Mintegral 10.90%, ironSource 0.54% and TikTok 0.27%. Facebook, Instagram, Audience Network and Threads return no plans in the monitored window — every other account we have broken down in the last three weeks ran Meta as its primary or co-primary surface. Whether Meta inventory shows up in a later snapshot remains to be seen; in this window the buying sits on the networks that monetise casual games.

That choice sets the creative format before a single asset is made. Meta rewards a hook in the first two seconds and punishes anything that looks like an ad. AdMob and the ad networks sit inside other apps, where the viewer is already in a game and the ad has to win attention mid-session rather than compete for it in a feed. The formats those networks reward — interactive playables, short loopable video, static end-cards — are the formats BirdFly actually ships.

The one plan per creative habit

367 plans and 308 creatives is not an account built by cloning. Compare it with Meowdoku, which ran 41,002 plans against 15,210 creatives in a single month, or Fruit Days, which spread 2,460 plans over 1,084 creatives. BirdFly sits at 1.2 plans per creative. Every asset gets its own plan; almost nothing gets cloned.

The reinvestment number confirms it from the other side. Only 15 of 308 creatives carry a re-run flag — 4.9%, against 53% for Fruit Days and 49% for Cozy Florist. The account is not running a bench of proven winners; it is continuously refreshing the library.

Two production lines inside one account

Sort the window's heaviest creatives by channel and the account splits cleanly in half.

The AdMob line is static-led. Its four heaviest assets are images at 432,491, 406,600, 338,378 and 320,379 estimated exposures, alongside video at 388,544 and 372,540. These are the units that sit inside other apps as banners and interstitials; they are cheap to produce and cheap to replace, and the creative copy is a single line — “A colorful world full of adorable birds is waiting for you.”

The network line is playable-led. The heaviest asset in the entire window is a Unity playable at 631,150 estimated exposure, live from October 5 to November 8 — thirty-five days of continuous delivery from one interactive build. AppLovin playables follow at 546,088, 529,037 and 518,527, with Unity playables at 530,329. Five of the twelve heaviest creatives are playables, and every one of them sits on a network rather than on Meta or YouTube.

The AdMob static line adds one more detail: those images are the only assets in the account with real similarity clusters. The heaviest re-run image carries fifteen near-duplicate siblings, the next five, then four and three. The playables, by contrast, ship as single builds with no clones. The cheap format is the one being multiplied; the expensive format ships once and runs for a month.

Inside the two lines: the creatives doing the work

The gameplay is observation, not reaction: a board full of birds in bubbles, four nest quotas at the top, and the player spotting three of the same species. Both sixty-second video leads frame the mechanic as a challenge instead of a demo — the familiar “only sharp eyes can win” bait — except here the challenge shown is the real game.

The playable line ships the same promise in interactive form. The heaviest Unity playable opens straight into the core loop with no tutorial copy: four nest quotas (0/2, 0/2, 0/3, 0/3), a 180-move bar, a feather-collection counter at 0/120, and a single green PLAY NOW button. Nothing is hidden and nothing is exaggerated — the playable is the game's first level with an install prompt bolted on, which is exactly why it can run for thirty-five days without fatigue.

Thailand leads a wide buy, not a deep one

Thailand covers 60.76% of plans — 223 of 367 — and is the account's largest market by a wide margin. Japan follows at 48.77%, Brazil 46.87%, Germany 46.05%, France 44.41%, Spain 43.60%, Mexico 41.14%, the United States 40.87%, Singapore 34.60% and Pakistan 33.79%. More than a hundred markets show up in the targeting data.

The exclusivity counts tell you what kind of buy this is. Thailand has 223 plans touching it and only 9 that touch nothing else. Japan: 179 and 7. Brazil: 172 and none. Pakistan: 124 and none. The United States is the outlier with 13 exclusive plans out of 150, and Singapore 11 out of 127 — the only two markets with any locally built plans.

Set against Fruit Days, the contrast is stark. Fruit Days parked 303 of its 306 Thailand plans as Thailand-only — a market-level batch with local creative. BirdFly parks 9 of 223. Same country at the top of both accounts, two different plays: Fruit Days built for Thailand, BirdFly swept Thailand in with broad targeting.

Always-on rotation, not a launch spike

Daily activity tells a different story from the plan-per-creative ratio. Active creatives opened the window at 145 on October 10, slid to a floor of 85 on October 14, then rebuilt through the second half of the month to 114 by October 23. November opened stronger: 116 creatives on the 3rd, a window peak of 125 on the 4th, and 119 as recently as November 8.

That is neither launch decay nor a burst. It is an always-on account holding roughly a hundred live creatives, with new ones rotating in — 19 new creatives on November 3, 18 on October 22, 14 on October 28 — as old ones retire. The spend goes to constant replacement rather than to scaling winners.

What the channel choice trades away

If the window's channel mix is the account's real structure — and two months of delivery suggests it is — the trade-off is worth naming. Meta is where paid social buys reach people who are not already inside a game. Leaving it out of the rotation, at least for now, gives up two things: the social graph — shares, comments, the organic tail a good hook can generate — and the ability to test a hook as a piece of content rather than as an interruption.

What it gains is a cleaner signal. AdMob and the other networks buy impressions inside gaming contexts, where the viewer already knows what is being offered. For a sorting game with no story and no characters to sell, that context is nearly the whole pitch: the unit shows the mechanic, the mechanic is the product. There is no fantasy to sell and therefore no gap between the ad and the game — the very gap, incidentally, that got four studios' ads banned by the UK regulator last week.

The scale is modest: 31.0M in estimated lifetime creative exposure across 367 plans, against 27.9M for Fruit Days on a much narrower market footprint. But the download picture is worth noting — Google Play reports a 100,000+ public label and an internal snapshot of 271,634, six weeks after launch, on a paid strategy that, in the monitored window, ran entirely on the ad-network side of the market.

Three things this account says about channel choice

First, the channel decision is a creative decision. A playable at 631k exposure for thirty-five days is only possible because the buyer chose networks that accept interactive units at scale. On Meta, playables are not a supported primary format; the equivalent budget would have gone into short video. The channel map and the format mix are the same decision made once.

Second, low reinvestment is a strategy, not an oversight. 4.9% reuse means the account pays full production cost every cycle in exchange for freshness. Where a reuse-heavy account like Fruit Days re-runs half its library each month to protect tested winners, BirdFly keeps rotating in new production. Both work; you cannot optimise for both at once.

Third, coverage is not commitment. Thailand at 60.8% of plans reads like a Thailand bet until you see 9 exclusive plans out of 223. Breadth at the top of a market list is not the same as depth in a market, and the exclusivity column is the only place that distinction shows up.

Questions we get about this one

Is BirdFly on Meta at all?

Not in the October 10 – November 9 monitored window, where Facebook, Instagram, Audience Network and Threads return no ad plans for this title. Seven other networks do. We report what the monitored set contains; absence of a plan in the window is not proof that no Meta placement has ever run, and later snapshots may differ.

Why does one account use both playables and static images?

They serve different units. Playables carry the interaction inside network placements and absorb the largest exposures (470–630k each). Static images are the cheap AdMob inventory that fills banners and interstitials, and they are the only assets in the account with large clone families.

Is Thailand the main market?

Thailand has the highest plan coverage at 60.76%, but only 9 of its 223 plans target Thailand alone. It leads a wide buy rather than anchoring a single-market build.

What does a 4.9% reinvestment rate mean?

Only 15 of 308 creatives carry a re-run flag in the window. The account is replacing creative rather than re-running proven winners, which keeps the library fresh but raises monthly production cost relative to reuse-heavy accounts.

How much exposure does the whole book carry?

31.0M estimated lifetime exposure across 308 creatives and 367 plans — modest in absolute terms, and roughly comparable to Fruit Days' 27.9M on a much narrower market book.