Gossip Harbor Ads: Four Years In, Still Shipping New Ads

Gossip Harbor turned four in September 2026 and spent the whole month feeding its ad engine. Across 30 days the account put 6,542 creatives and 7,860 ad plans into delivery, added new material on all 30 of them, and skewed its buying almost entirely toward AdMob and AppLovin while the Meta family handled a small, oddly specific job.

Data scope

Gossip Harbor: Merge & Story is a merge-and-story title from Microfun Limited with a release date of September 9, 2022, so it turned four during the window this piece covers. Creative, ad plan, market and exposure figures come from CreativeScope and cover September 1 to 30, 2026 unless stated otherwise. They describe two store records: the main record used throughout, and a secondary record reported separately in the notes and never added to the first.

An ad plan is one buying configuration: a channel, a set of target markets and a language. It is a delivery structure, not a money figure, and none of the counts below measures what the campaign bought or what came back. The practice these numbers describe is user acquisition, often shortened to UA by practitioners, meaning the paid install-driving side of growing a mobile game.

A month that never had a zero day

The defining fact of the window is the consistency of supply. Over 30 days the account added 2,895 new creatives and 3,321 new ad plans, and new material appeared on every single day of the month. The daily average is about 96 new creatives; the quietest day, September 2, still added 41. Peaks cluster in two bands: September 23 with 190 new creatives and 209 new plans, September 24 with 171 and 184, and, earlier in the month, September 7 with 163 and 165 and September 4 with 147 and 180.

That steady flow is visible in the active base. Daily active ad plans stay inside a narrow band all month, between 4,299 on September 8 and 4,574 on September 24, a spread of about 6%. Four-year-old titles usually show this shape for a reason: the creative pipeline is a content calendar, and the story episodes that ship every week need matching ad supply. Reinvestment is part of the same machine, with 525 creatives returning to delivery after a pause during the month.

Two engines: AdMob and AppLovin

The channel structure is unusually lopsided for a title this size. AdMob covers 3,487 plans (44.36%) and AppLovin 3,008 (38.27%), so two networks account for 82.6% of the month's buying. Unity Ads adds 401 plans (5.10%), Vungle 344 (4.38%) and YouTube 145 (1.84%).

The Meta family sits far below its usual share for a casual title: Facebook, Instagram, Audience Network and Threads each cover the same 466 plans (5.93%), with Messenger at 360 (4.58%). Read together with the creative formats below, that small Meta footprint looks like a deliberate assignment rather than neglect. The short merge loops run where inventory is cheap and automated; the longer localized story videos, the ones that carry character drama and promise copy, run on the Meta surfaces where a viewer will sit through a minute of narrative.

Markets: shared coverage with two exclusive blades

Coverage is broad and mostly shared. France leads with 2,926 plans (37.23%), then Switzerland 2,899 (36.88%), Singapore 2,763 (35.15%), Canada 2,399 (30.52%), Japan 2,393 (30.45%), Germany 2,284 (29.06%), Spain 2,269 (28.87%), Australia 2,189 (27.85%), the United States 1,827 (23.24%) and Thailand 1,809 (23.02%). Most of those are multi-market plans reused across groups of countries.

Where the account pays for separation, it does so in two places. Japan keeps 502 exclusive plans of the 2,393 covering it, the largest single-market block in the window by a wide margin, and Australia holds 263 of 2,189. Every other market in the top ten keeps an exclusive layer under 7% of its own coverage: Spain 31 of 2,269, the United States 72 of 1,827, Thailand 42 of 1,809. Two exclusive blades inside an otherwise shared structure is a clear pattern, and the two markets holding them are also the ones with localized long-form video in the spotlight set below.

Four creatives: loops, stories and one promise

The four spotlight creatives below are drawn from the window's top files by estimated exposure, chosen so that no two share a storyboard. Platform, dates and estimated exposure are CreativeScope metadata; the scene descriptions are observations of the downloaded files themselves. Two of the four are six-second merge loops, and two are long localized narrative videos.

What four years of buying can and cannot tell you

The reading in one sentence: this account has industrialised creative supply and split the work by network, with AdMob and AppLovin carrying the merge loops at scale and a small Meta footprint reserved for localized narrative formats that make a promise about the experience. For a UA team, the transferable move is to check whether your own channel split matches your format split, because here they line up almost exactly.

Alternative explanations deserve their own paragraph. The two-network concentration may reflect how the account's mediation and buying are set up rather than a considered preference, and nothing here shows whether the Meta surfaces were tested more heavily at another time. The Japan and Australia exclusive blocks may reflect store-account structure, local publishing arrangements or language needs rather than a deliberate market bet.

The boundaries hold regardless. An exclusive plan shows targeting intent, never results; a market appearing in a target list does not imply a dedicated budget; estimated exposure is not a platform-measured impression count; and nothing in plan counts, market coverage or exposure estimates supports conclusions about cost per install, return on ad investment, or how any of these creatives actually performed.

Frequently asked

What counts as an ad plan in this article?

One buying configuration: a channel, a set of target markets and a language. In September 2026, the main Gossip Harbor record ran 7,860 of them. Plans measure delivery structure, not money.

Does 6,542 creatives in a month mean a huge budget?

The data cannot say. Creative and plan counts describe activity, and nothing in them translates into budgets, cost per install or revenue. High supply volume is compatible with both large and small budgets.

Why is the Meta share so low for a casual game?

That is what the tracked plans show: each Meta surface covers 466 of 7,860 plans (5.93%), while AdMob covers 44.36% and AppLovin 38.27%. The account's localized long-form videos do run on Meta, which suggests the surfaces are assigned to a specific format rather than dropped.

What does an exclusive plan actually prove?

Targeting intent. A Japan-exclusive plan lists Japan only, so the account paid for market separation there; it says nothing about how those plans performed.

Why report two store records separately?

Because they are separate advertiser records with different release dates and different scale. Adding them would inflate every number in the article, so the secondary record appears once, in these notes, and is excluded from all counts.

Where does the data come from?

From CreativeScope, covering September 1 to 30, 2026, for Gossip Harbor: Merge & Story. Scene descriptions come from the advertiser's own creative files, and no external numbers are mixed with CreativeScope figures.

Written by Maya Chen for CreativeScope. Creative descriptions come from the advertiser's own video creatives. Creative and ad plan counts are delivery-activity measures, not budgets or performance, and this article makes no claims about conversion, retention or commercial outcome. Cover image cropped from advertiser material. All trademarks belong to their owners.