Merge Ads Are Getting a Plot, Told by Characters Not Actors
Story-led merge ads took 43% of estimated exposure in 60 days while live-action stayed at 3.6%. What 253 creatives say about where the genre is heading.
Daniel Brooks
Creative Strategy Analyst
Published
Updated
We opened 253 merge-2 creatives frame by frame over sixty days. The budget is going to storyboards - and almost none of it is filmed with real people.
Where the genre standsMerge-2 is no longer a young genre, and in the ad libraries it does not behave like one. In our own records, ten leading titles put 2,467 unique creatives into circulation in a single sixty-day window - 1,723 of them video - and 71.3% of that video sits in the same slot: a vertical cut between 30 and 59 seconds. That is a mature format with a house style, not a category still experimenting with shape.
The buying side is just as settled. No market takes a decisive share of it: the top country on Android carries 3.7% of attributed ad plans and the next four sit between 2.9% and 3.4%. The mechanic that made the genre legible - put two items together, watch them become one - is still the default shot, and it runs across most of the pool. When every title buys the same format into the same spread of markets, the one variable left to move is what the ad actually says.
Eleven seconds into a Travel Town ad, a mother and a child stand in the snow outside the house they have just lost. Nothing has been merged yet. The next forty-seven seconds are about getting them back indoors - heat, a window, a bed - and the ad never once stops to sell a board.
Merge-2 has always sold itself with the satisfying click of two items becoming one, and that shot still fills the pool. So we opened 253 creatives frame by frame across ten titles and sixty days to check two things: whether the genre has really turned narrative, and whether the turn came with actors attached. It has. It did not.
Key Takeaways
- Story is winning the money, not the count. Story-led directions (crisis-and-repair, plot-twist suspense, fashion makeover) are 79 of 253 reviewed creatives but 43.0% of estimated exposure. Among the 110 creatives first seen inside the window, the exposure share rises to 49.2%.
- The plots are family, relationships and self-image. A mother and child losing their house in the snow, a bride with a secret, a failed school outfit before a makeover. The genre's biggest single exposure bucket is crisis-and-repair at 22.3%.
- It is not a live-action turn. Nine creatives - 3.6% - carry any live-action element, and one Gossip Harbor ad holds 87% of that exposure. Where live action appears, it is a 3-to-4-second hand or phone opening, then the ad is animation.
- Story and mechanics live on different networks. Google-family inventory (AdMob, YouTube) takes 87.0% of mechanic-demo exposure, with Meta getting just 1.8%; on the story side, Meta's family carries 41.2%.
- 30 to 59 seconds is the master format. 1,229 of 1,723 videos (71.3%) sit in that band, 85.6% of videos with orientation data are portrait, and story creatives run longer (median 58.3s) than mechanic demos (50.9s).
- Supply leans female-led; demand is unproven. The money-carrying plots put a woman at the center of the ad. That is a statement about what studios animate, not about who installs - the data here cannot resolve audience gender.
Data Scope
Window: 13 August to 12 October 2026, queried 13 October 2026. All figures come from CreativeScope ad records for ten merge-2 titles - Microfun, Century Games, Moon Active, Happibits, Metacore, Cedar Games and VoyagerOne, each queried per store: Gossip Harbor: Merge & Story, Seaside Escape: Merge & Story, Tasty Travels: Merge Game, Travel Town, Hollywood Merge, Merge Cooking®, Foodstars: Merge & Cook, Mystery Town: Merge Games, Hotel Legacy: Merge Game and Merge Mansion.
The ad-plan pool holds 51,795 Android and 13,169 iOS query lines; after deduplication by media address it contains 2,467 unique creatives (1,723 video, 196 image, 548 playable), of which 253 were reviewed frame by frame - 110 of them first seen inside the window. Every direction label and storyboard in this article comes from that reviewed layer.
An ad plan is one buying configuration: one network, one market set, one language. A creative is the asset itself. Android and iOS are separate query lines and are never added together. Exposure is a lifecycle estimated-exposure value attached to a creative - a delivery-heat estimate, not spend, revenue or measured views. Country figures are target-list coverage, not exclusivity.
What the ten-title pool covers
The pool is the ten merge-2 titles with the largest in-window creative volume that we track across both stores. It is a volume-based working set, not a category ranking, and it skews toward titles that ship creatives constantly - which is the point, because the question here is what the genre is making right now.
| Title | Studio | Window creatives | Reviewed directions (top) | |
|---|---|---|---|---|
 | Foodstars: Merge & Cook | Happibits Game | Android 12,694iOS 1,182 | Food serving, board orders, production |
 | Mystery Town: Merge Games | Cedar Games Studio | Android 9,099iOS 2,342 | Reward offers, tabletop demos, makeovers |
 | Gossip Harbor: Merge & Story | Microfun Limited | Android 9,167iOS 1,232 | Board orders, crisis repairs, story suspense |
 | Tasty Travels: Merge Game | Century Games | Android 5,654iOS 3,033 | Tabletop launching, container stacking |
 | Hollywood Merge | VoyagerOne | Android 5,745iOS 1,854 | Fashion makeovers (all 24 reviewed) |
 | Merge Cooking® | Happibits | Android 4,556iOS 1,398 | Board orders, production, hidden objects |
 | Seaside Escape: Merge & Story | Microfun Limited | Android 2,012iOS 413 | Board orders, crisis repairs |
 | Hotel Legacy: Merge Game | Century Games Innovation | Android 851iOS 993 | Production, tabletop demos |
 | Merge Mansion | Metacore Games Oy | Android 1,508iOS 289 | Story suspense, production |
 | Travel Town | Magmatic Games (Moon Active) | Android 509iOS 433 | Board orders, crisis repairs |
Story takes the money, mechanics take the count
Open 253 creatives and label each one's direction, then put every label into one of four families. Mechanic demonstrations - board orders, tabletop launches, container stacking, production lines, food serving, hidden-object boards - are the majority of what gets made: 136 creatives, 53.8% of the sample. Story-led work - crisis-and-repair arcs, plot-twist suspense, fashion makeovers - is 79 creatives, 31.2%.
Exposure tells the opposite story. The mechanic family holds 49.1% of estimated exposure, less than its sample share. The story family holds 43.0%, nearly twelve points above its sample share. The money-carrying creative in merge is, right now, more likely to be a narrative than a demo.
Drill into directions and the leader is not a demo at all. Crisis-and-repair stories hold 22.3% of estimated exposure, ahead of tabletop launch demos at 21.9% and board-order demos at 15.6%. Fashion makeovers, the second story direction, add another 14.2%. Taken together, the three story directions out-carry any two mechanic directions combined.
The window's fresh layer pushes the same way. Of the 110 reviewed creatives first seen inside the 60 days, story directions take 34.5% of the count and 49.2% of the exposure, against 28.7% and 41.9% for the older stock. New work is not drifting back toward demos; it is drifting further into plot.
The plots are family, relationships and self-image
What is the story, concretely? The crisis-repair arc opens on loss. In Travel Town's 59-second ad, a mother and child lose their shelter in the snow. Gossip Harbor's top Android story creative opens on an injured woman in a damaged house, and the merge board alternates with bed, window and stove repairs until the room visibly recovers.
The suspense arc trades on secrets. Merge Mansion's 36-second CG ad opens on a distressed bride at a wedding, cuts to a police car and a bedside, and closes on the line "Every family has its secrets." The makeover arc sells identity progression: Hollywood Merge's 68.7-second ad runs a dress-for-school choice that ends in a FAILED stamp, then walks through wedding, red-carpet and party scenes.
Notice who is inside these plots. The person who loses the house is a mother. The person with the secret is a bride and her grandmother. The person being remade before a date or a school day is a young woman. Across the story family's 79 creatives, the recurring premise is a woman in trouble, a woman being transformed, or a relationship under strain. Studios are not animating a generic hero; they are animating a female-led home-and-family story.
That is the closest the creative record gets to answering whether merge demand skews female. It is evidence about supply: what the category leaders behind those 2,467 creatives choose to produce and buy. What it cannot tell you is who converts on the other end - that question needs install and retention data, which ad libraries do not carry.
It is not a live-action turn
The other hypothesis going around merge-2 circles is that the genre is going live action, the way puzzle ads adopted on-camera presenters. In this pool, it has not happened. Nine of 253 reviewed creatives - 3.6% - carry any live-action element: a real hand, a real phone, an on-camera presenter, or a realistic-looking character in the opening seconds.
Their combined estimated exposure is 6.9% of the reviewed total, and 87% of that sits on a single Gossip Harbor iOS ad that opens on a live hand close-up and runs 46 seconds. Remove that one asset and the rest of the live-action set amounts to 0.9% of exposure.
Look at where live action does appear, and a pattern emerges: it is an opener, not the ad. Seaside Escape's snow-boot creative gives a real hand four seconds - touching the boot, breaking the snow - before the remaining 41 seconds play as game animation. Hotel Legacy uses realistic hotel characters for a 6-second front-desk opening, then moves to production management. The live element does the grabbing; the game does the selling.
Our previous pool used a stricter, on-camera-only definition and found five presenter openings; this window's looser element-level count finds nine. Both readings land in the same place - live action is a seasoning a few studios add to story openings, not a direction the genre is moving into.
Half a minute is the floor
Of the 2,467 creatives, 1,723 are video, 548 playable and 196 image. Video runtime clusters hard: 1,229 videos (71.3%) run 30 to 59 seconds, and only 140 (8.1%) run under 10 seconds. Story creatives run longer than demos - a median of 58.3 seconds against 50.9 - and the brand-and-seasonal family is the only one comfortable in short form, with a 15.2-second median. Where orientation is reported, 1,474 of 1,722 videos (85.6%) are portrait. Playables are not a side channel: 548 creatives, at 23.9% of the iOS line and 20.7% of the Android line.
Story and mechanics live on different networks
The split that matters for planning is which inventory carries which kind of creative. Sum estimated exposure by network family inside each direction family, and the two rows barely overlap. Mechanic demos are a Google-network business: AdMob and YouTube together carry 87.0% of mechanic exposure, Meta's placements carry 1.8%. Story-driven creatives are far more balanced: 48.2% Google, 41.2% Meta, 9.0% AppLovin.
Plan-level attribution agrees. On Android, AdMob holds 20.8% of attributed plans, and the Meta family taken together - Facebook 14.8%, Instagram 14.6%, Audience Network 12.7%, Threads 11.9%, Messenger 7.6% - adds to roughly 62%. On iOS the Meta family alone reaches about 84% of attributed plans. Attribution is multi-touch, so shares add past 100%; the direction-level exposure split above is the cleaner read on where each creative type actually delivers.
Markets, meanwhile, are flat. The top country on Android is Germany at 3.7% of attributed plans, followed by Japan 3.4%, France 3.3%, Singapore 3.1% and the UK 2.9%; on iOS, the UK 3.0%, Singapore 2.9% and the US 2.8% lead. No single market holds 4% of the pool. A merge buy is a wide, shallow coverage map, which is also why the creative has to travel: a plot that needs language-free storytelling - loss, repair, transformation - scales across that map more cheaply than dialogue does.
What this does not prove
- Creative volume is supply, not demand. Studios chose to animate female-led home-and-family plots; that says nothing direct about conversion, retention or payer mix. The female-lean is an observed production pattern, not an audience measurement.
- Exposure is not spend. Estimated exposure is a delivery-heat value attached to the creative across its whole lifecycle. It ranks assets; it does not price them.
- The reviewed layer carries old long-runners. 110 of the 253 reviewed creatives were first seen inside the window; the rest are older assets still buying. Trend statements here are marginal signals, not reversals.
- Attribution is multi-touch. Network shares sum past 100% by design, so a "share of plans" is presence, not an exclusive allocation.
- Two stores, two lines. Android and iOS counts are never summed; pool-level statements use the deduplicated creative set.
What a UA team can do with this
- Cut for the destination. A 30-to-60-second crisis-and-repair plot has a home on Meta's placements; a 6-second board demo has a home on Google's network. Mechanic demos run 87.0% on Google-side inventory against 1.8% on Meta - re-cutting one into the other wastes both.
- Build the plot from a female-led premise if your title supports one. Crisis-and-repair is the single largest exposure direction at 22.3%, and its working formula - loss, choice, repair, visible recovery - is language-free and travels across the genre's flat market map.
- Use live action as a 3-second opener at most. Nine creatives carry it, one carries the exposure. If a test needs a face, spend it on the first seconds, not the whole runtime.
- Treat 9:16, 30-to-59 seconds as the master spec. 71.3% of video sits in that band. Short-form is a separate lane owned by brand and seasonal assets, not a re-cut of the master.
How this report was built - and how to run it yourself
Where the data comes from. Every figure in this piece is drawn from CreativeScope: creative records across 15 ad networks and 53 countries and regions, more than 3.5 million new ad plans indexed per month, refreshed hourly. Nothing here is estimated from a panel or a survey - it is what the ad libraries hold.
How the analysis was produced. The analysis layer was generated with AI assistance. An AI agent pulled each title's creative pool per store, deduplicated it by media address, opened the high-signal creatives frame by frame, and wrote the direction, hook and storyboard labels used above. A human reviewer then checked every label and every quoted storyboard against the frames before publication. The numbers are CreativeScope's; the reading of them is a machine pass plus an editorial pass, which is also why the "What this does not prove" section exists.
The tool the agent used. It did not click through a dashboard. It called the CreativeScope MCP server, which exposes the same creative-intelligence toolset we run internally to any AI client that speaks MCP: keyword and filter search, advertiser profiles, material detail with playback, hook analysis, similar-creative lookup and store first-seen tracking. If your team already runs an AI assistant, this is the shortest path from a question about a competitor to a structured answer with sources.
The keywords behind this report. If you want to reproduce the query, these are the terms and filters it used:
- Category and query terms: merge, merge-2, merge game, merge & story, merge puzzle, plus the ten store titles listed above.
- Filters: 60-day window (13 Aug - 12 Oct 2026), platform Android and iOS as separate lines, formats video / image / playable, all markets, ranked by estimated exposure.
- Labels applied on review: creative direction (crisis-and-repair, plot-twist suspense, fashion makeover, mechanic demo, reward offer, brand & seasonal), hook in the opening three seconds, runtime band, orientation, live-action flag.
Run it on your own titles. Paste the prompt below into your AI client with the CreativeScope MCP server connected, swap the title, and you get the same structure back:
Two ways in: connect the CreativeScope MCP server to the agent you already use, or create a free CreativeScope account and run the same search in the browser. Either route reads the same library this article was written from.
Related Reading
- Merge Game Ads: From Trash to Treasure, Two Ways - the previous merge-2 pool, 12 titles and 445 opened creatives, with the serving/renovation axes spelled out.
- Merge puzzle ads under one title - what happens when a merge buy concentrates its identity.
- Gossip Harbor: drama before gameplay - the crisis-repair formula at single-title depth.
FAQ
What counts as a story-led creative? One whose reviewed direction is crisis-and-repair storytelling, plot-twist suspense or fashion makeover - the asset advances or sells a narrative with characters, not a mechanic demonstration. Reward-copy and brand-scene creatives are kept in separate families.
Is estimated exposure the same as spend? No. It is a delivery-heat estimate attached to a creative across its lifecycle. We use it to rank creative weight, not to price media.
Why review 253 of 2,467 creatives? The full pool answers structural questions - formats, runtimes, orientation, network presence - at complete coverage. Direction labels and storyboards need frame-by-frame viewing, which we spend on the high-signal layer; 172 of the 253 reviewed creatives are flagged high-signal.
Data Notes
Window. 13 August to 12 October 2026; queried 13 October 2026.
Store lines. Android and iOS queried separately; counts are reported per store and never summed.
Exposure. Lifecycle estimated exposure per creative; sums are over the 253 reviewed creatives unless stated. Percentages may not sum to 100 due to rounding.
Direction labels. Twelve directions from frame-by-frame review of 253 creatives; storyboards quoted in this article are from the reviewed storyboard set of 20.
Live-action flag. Any live hand, phone, presenter or realistic-looking character in the opening sequence; looser than the on-camera-presenter definition used in our previous pool.
Countries. Target-list coverage per ad plan, not exclusivity; coverage is not spend.
Attribution. Multi-touch across networks; shares add past 100%.
Get this live in your agent
These rankings come from a single MCP tool. Free key, ask in plain English.
Full genre × market grid?
Team plan unlocks every genre and market cut, plus AI teardowns.