The short answer
Yes, optimize in cheaper markets first. In fact that has already started: the September GA4 data shows Meta installs in Mexico and Brazil. Those players left fast, with 21% back on day 1 and about 5% by day 7. So the first job is keeping players through their first session, not buying more of them. The usual cheap test countries in Asia are also closed to this game on Meta.
No US scale yet
A US Android install for a casual puzzle game costs about $1.20 to $3.50. In September, 4 in 5 paid players were gone by day 2 and nobody bought anything, so US installs would mostly be money spent on players who leave.
Test creative in Brazil and LATAM
Keep testing where September ran, at a fraction of the US price. Brazil players played twice as long as Mexico players (34 against 17 minutes each), so Brazil leads round 1.
Treat it as social casino
The in-board slot reels, "BIG WIN" moments and coins likely put SlotoMatch under Meta's social casino rules. Ads must target 18+, and Meta bans this category in the Philippines, India, Indonesia, Vietnam, Thailand, Malaysia and 13 more countries.
Fix the first 12 levels
Only 34% of new players reached level 10, although levels 1 to 10 are won 97% of the time. Players are leaving because the opening is slow, not because it is hard. Rewarded ads already exist; they need an ad network that pays and a second placement.
What the September data shows
The GA4 export in the asset library covers 17 to 25 September 2026: app version 1.7 and up, the team in Israel excluded, the last day partial. It is a small sample of nine days and 348 players, so read the numbers as direction, not precision. It has no spend, so cost per install is still unknown.
Who came and from where
| Signal | Value | What it means |
|---|---|---|
| Active players | 348 | 295 installs in the window |
| From paid Meta | 73% | 305 players, 245 of them new. Direct 21%, Play Store search 5% |
| Mexico | 201 players | 187 new; 17 minutes played each |
| Brazil | 115 players | 95 new; 34 minutes played each, twice Mexico |
| United States | 4 players | No US data yet |
| Uninstalls | 66% of installs | 196 of 295 uninstalled inside the nine days |
How long they stayed
| Players | Count | Day 1 | Day 2 | Day 3 | Day 7 |
|---|---|---|---|---|---|
| Paid Meta | 246 | 21% | 12% | 5% | 5% |
| All players | 262 | 27% | 14% | 6% | 5% |
| Match-3 median (benchmark) | – | ≈ 32% | – | – | ≈ 14% |
Day 7 rests on the 17 September cohort alone (42 players). Direct players returned at 70% on day 1, but most of them already knew the game (only 47 of 89 were new), so they are not a benchmark.
Where they left
| Level reached | 5 | 10 | 12 | 20 | 30 | 50 | 100 |
|---|---|---|---|---|---|---|---|
| Share of level-1 players | 59% | 34% | 26% | 20% | 13% | 7% | 2.5% |
- The opening loses players without being hard. Levels 1 to 10 are won 97% of the time. The biggest early drops come after levels 2, 5 and 11 (16 to 18% each), and 9 of the first 11 levels start with a tutorial.
- Some levels are too easy: players finish levels 3, 6, 18, 20 and 25 with 9 or more moves to spare.
- Later spikes: level 47 is won 35% of the time, and levels 67 to 80 and 92 to 104 mostly sit between 14% and 60%. Few players get that far, so these matter less than the opening.
What the game already has
- Rewarded ads exist at the out-of-moves popup: 24 players watched 117 ads. At that popup 30% pay coins for moves, 12% watch an ad and 57% give up.
- No real-money purchases: 36 players (10%) opened the shop and none bought.
- Retention features are built: chests, missions, story upgrades (91 players), the Captain's Hook claw goal (55 players) and a daily intro that only half of players finish. What's missing is a schedule of events that gives players a reason to come back.
Things in the brief that need checking
- The store says the game contains ads and in-app purchases. The data shows ad events but no purchases. Confirm the ads are served by a network that pays, and that purchases work end to end.
- The brief names the crew Otto, Red, Poly and Rex. The store spells Poli, and the shark that hosts the Captain's Hook bonus game has no name on the store. Ads need one fixed set of names.
- The store rates the game Everyone. If its rating answers left out the simulated slot mechanics, the rating may change once they are declared. That is worth checking before we advertise to 35 to 65 year olds with slot imagery.
- Meta's public Ad Library shows no SlotoMatch or Romzy ads, because it keeps finished ads only if they ran in the EU. The September campaign's spend, cost per install and creatives are in Ads Manager; an export would add them.
Which cheaper markets
Cheap markets answer creative questions well: which hook stops the scroll, which video gets installs, which players finish the tutorial. They do not tell us how a 55-year-old American will play or spend. So the plan uses cheap markets to pick creative and a small US and Canada test to confirm before scaling.
| Market | Expected Android CPI | Role | Notes |
|---|---|---|---|
| Brazil | $0.15–0.60 | Round 1 lead | Ran in September: 115 players, 34 minutes each. Needs Portuguese ad text and store listing. |
| Mexico, Colombia, Argentina | $0.15–0.60 | Creative test | Mexico ran in September: 201 players, 17 minutes each. Needs Spanish ad text and store listing. Benchmark is for LATAM overall. |
| South Africa, Nigeria, Kenya | Unknown, test | Round 2 cell | English-speaking, so the same ads run unchanged. Untested and no published benchmark, so it follows the markets we have data for. |
| Canada | Tier 1 | Proxy test | Closest behaviour to the US, slightly cheaper and smaller. |
| United States | $1.20–3.50 | Proxy, then scale | Target market. Small cohort first, scale only after the gates below. |
| Philippines, India, Indonesia, Vietnam, Thailand, Malaysia | n/a | Not allowed | Meta bans social casino and free-to-play casino ads here. Also banned: Azerbaijan, Bangladesh, Cambodia, Egypt, Hong Kong, Korea, Kyrgyzstan, Mongolia, Myanmar, Pakistan, Saudi Arabia, Singapore, Taiwan. |
CPI ranges are vendor benchmarks for casual and puzzle games on Android and are not Meta-specific; Meta often runs higher than Google App Campaigns. Treat them as starting expectations, not targets.
The plan, in gated phases
Each phase ends with a check. We only move on when the numbers pass it, and every launch and budget change goes to you for approval first, as the brief requires.
Keep players through the first session
- Faster opening: fold the tutorials on levels 1 to 11 into fewer, shorter ones; let the board reels fire by level 2, since they are the game's hook; tighten the too-easy levels 3 and 6. Target: 50% of new players reach level 10, up from 34%.
- Reasons to return: the daily intro, missions and chests exist. Add a 7-day login streak, one weekly event or leaderboard, and push notifications for both. Find out why half of players skip the daily intro.
- Tracking: GA4 already records levels, tutorials and ads. Add the Meta SDK with activate app, tutorial complete and level achieved (3, 10 and 20) firing in Events Manager, and pass campaign and ad names into GA4 so retention can be read per ad.
- Monetization: rewarded ads already run at the out-of-moves popup. Confirm a paying ad network, add a second placement (double chest rewards), and test a first low-price coin pack, since 10% open the shop and nobody buys.
- Bonus links: the brief asks for bonus-link posts on the Page. The game needs a deep link that grants the reward before those posts can work.
- Store listing: short description and first two screenshots aimed at "match 3 pirate puzzle", Spanish and Portuguese store text for Phase 1.
Creative test in cheap markets
- App install campaigns on Android, ages 18+ (required) with a 35–65+ cell and a broad 18+ cell to see whether the age limit helps.
- ABO for the test so each creative gets a fair share: 8 to 12 creatives in ad sets of 3 to 4, $20–30 a day per ad set.
- Two clusters: Brazil (Portuguese) and Spanish-speaking LATAM (Mexico, Colombia, Argentina). English-speaking Africa joins as a cell in round 2. Advantage+ placements.
- Read day 1 retention for each ad set in GA4 as well as cost per install, so we keep the ads that bring players who stay.
- Kill an ad after about 3,000 impressions if its link CTR is under 0.7%. Move winners to the next round and add 3 to 4 new concepts each week.
US and Canada proxy cohort
- The 3 best creatives from Phase 1, US and Canada, $50–100 a day, 35–65+ against broad.
- Optimize for installs until an app event reaches about 50 a week, then switch optimization to level 10 achieved.
- Read cost per install, day 1 and day 7 retention and rewarded-video revenue per daily player in the real target audience.
US scale
- Move winners to CBO or Advantage+ app campaigns. Raise budgets by about 20% every three days while CPI holds.
- Build lookalikes from players who reached level 20 or watched rewarded video, once there are 1,000 or more of them.
- Keep testing 3 to 4 new creatives a week, because install costs rise as an ad wears out.
Creative concepts from what we already have
Short vertical gameplay videos of 10 to 20 seconds, the first second showing the board. Each concept uses art from the store listing and the client's key art.
Copy rules for every ad
- Coins, moves and rewards only. Never "win money", "cash", "jackpot payout" or anything suggesting real-money prizes.
- Keep the store's line visible on the Page and store: for entertainment only, and success here does not mean success at real-money gambling.
- Target 18+ in every ad set, even organic boosts.
Page and community, at low cost
| Channel | What to do | Cadence |
|---|---|---|
| Facebook Page | Gameplay clips, crew moments, new levels and events. Bonus-link posts once the deep link exists. | 3 posts a week, daily bonus post later |
| Facebook groups | Join US groups for match-3 and free mobile games as the Page. Follow each group's rules, share only where promotion is allowed, answer questions, never post the same message across groups. | 2–3 groups a week |
| Reels and YouTube Shorts | The same short videos as the ads, posted organically. Costs nothing and shows which hooks people watch. | 2 a week |
| Google Play | Reply to every review, refresh screenshots after each test round, and run store listing experiments on the first screenshot. | Weekly |
| Google App Campaigns | For an Android-only game, worth a small parallel test; installs are often cheaper than on Meta. | After Phase 1 |
Scorecard
The numbers each phase is judged on, next to what paid players did in September. Day 1 and day 7 targets sit near the published medians for match-3 games, which are around 32% and 14%.
| Metric | September (paid Meta) | Phase 1 (cheap markets) | Phase 2 (US and Canada) |
|---|---|---|---|
| Link CTR | Not in GA4 | ≥ 1.0% | ≥ 1.0% |
| Installs per 1,000 impressions | Not in GA4 | ≥ 5 | ≥ 3 |
| Cost per install | Not in GA4 | ≤ $0.40 | ≤ $2.50 |
| Tutorial completion | 96–100% per level | ≥ 70% | ≥ 75% |
| Day 1 retention | 21% | ≥ 30% | ≥ 35% |
| Day 7 retention | 5% | ≥ 10% | ≥ 12% |
| Share of new players reaching level 10 | 34% | ≥ 50% | ≥ 50% |
| Rewarded video views per daily player | ≈ 0.2 | Measure | ≥ 2 |
These are proposals to agree with you before launch, not promises. The September baseline comes from nine days and a small sample.
Access and approvals
- As the brief sets out, all assets stay owned by the game owner. We need Advertiser access to the ad account (no billing or admin), Content and Community access to the Page, and view access to Events Manager for the app.
- Every new campaign, strategy change and budget change goes to you for written approval before it runs.
- Open questions for you: is there an iOS build planned; which languages does the game support; which analytics SDK is in the build; and when can rewarded video ship?
Sources
- SlotoMatch GA4 export, 17–25 Sep 2026 (property 536415198, app 1.7+, team excluded), in the asset library.
- Google Play listing, com.romzygames.slotomatch, checked 25 Oct 2026.
- Meta Ad Library, exact-phrase searches for "SlotoMatch" and "Romzy", all countries, 27 Oct 2026: facebook.com/ads/library.
- Meta Advertising Standards, Social casino games and Online gambling and games.
- CPI ranges: Cost Per Install Benchmarks 2026, Mobile Game CPI Benchmarks 2026, Playio CPI benchmarks, Business of Apps CPI rates.
- Retention: Segwise retention benchmarks 2026, Playio D1/D7/D30 benchmarks, GameAnalytics match-3 benchmark.