Level 1

Coins: 0. Open the shop

Playable game: stack ten growth levers. Press Space or Enter, click or tap to drop the sliding block on the tower.

d7tec, mobile growth studio

We scale mobile games.Strategy. Creative. User Acquisition. Analytics.

From soft launch to profitable scale.

Start scaling →See our results →
Skip the game

Stack to scale. Press Space or Enter, click or tap to drop each floor.

This playable needs WebGL. Skip the game and read the levels below.

Built for teams that own the number.

Studios we scale with

  • kitebird
  • HEXFORGEGAMES
  • TIDAL PIXEL
  • MoonmothSTUDIO
  • brickfox
  • Orbit Owl
  • SALTWAVEINTERACTIVE
  • pinecone play

Where we buy

  • Meta
  • Google Ads
  • TikTok
  • Apple Ads
  • Unity
  • Snapchat
  • AppLovin
  • Moloco
  • Mintegral
  • ironSource
  • Liftoff

Where is your game?

Four stages, one path to profitable scale. Pick yours and we start there.

  2. Stage 2: Soft launch

    • KPI validation
    • Creative testing
    • CPI/LTV testing
    • Monetization review
    I'm in soft launch →
  5. Boss level

    Profitable scale

What we actually do

Six pillars, run by one squad.

  • Growth Strategy

    GTM, market analysis, economics, forecasting.

  • User Acquisition

    Meta, Google, TikTok, AppLovin, Unity, Apple Ads, DSPs.

    • Meta
    • Google Ads
    • TikTok
    • AppLovin
    • Unity
    • Apple Ads

  • Creative Lab

    Video, UGC, playables, AI, concepts, testing.

  • App Store Growth

    ASO, CPP, store testing, localization.

  • Analytics & Attribution

    AppsFlyer / Adjust, SKAN, dashboards, cohort economics.

  • Monetization & Retention

    IAP, ads, LTV, retention, re-engagement.

The Scale Loop™

Stop buying agency hours. Buy a system.

  1. Measure
  2. Learn
  3. Create
  4. Acquire
  5. Optimize
  6. Scale

Data → hypothesis → creative → traffic → cohort → decision.

Results

Four games. Baseline, result and window on every card. Flip one for how it was measured and what failed.

Rarity follows peak monthly spend Legendary Epic Rare Uncommon

  • Stackopolis

    Rare

    Hybrid-casual14 weeks

    Monthly spend
    $40K → $280K/month
    D30 ROAS
    38% → 67%
    CPI
    −$0.74
    Revenue
    +214%

    Stackopolis How it was measured

    Genre
    Hybrid-casual city stacker
    Geos
    US, UK, DE, JP
    Window
    14 weeks, D30 cohorts
    Measured with
    MMP + SKAN CV schema + geo holdout
    What failed
    UGC testimonials ran at IPM 2.1. Puzzle players don't trust "real people".

    No. 001Tap to flip back

    See exactly how we did it →

  • Kraken Kitchen

    Epic

    Merge10 weeks

    Monthly spend
    $60K → $410K/month
    D30 ROAS
    19% → 31%
    CPI
    −$0.87
    Revenue
    +163%

    Kraken Kitchen How it was measured

    Genre
    Merge / cooking, IAP-heavy
    Geos
    Soft launch PH/CA/AU → global
    Window
    10 weeks, D30 cohorts
    Measured with
    MMP + predictive LTV (D3 → D180)
    What failed
    "Fake gameplay" ads: cheap installs, D1 22% vs 38%. Killed in week 2.

    No. 002Tap to flip back

    See exactly how we did it →

  • Tidebreakers

    Legendary

    Mid-core strategy16 weeks

    Monthly spend
    $90K → $520K/month
    D30 ROAS
    14% → 26%
    CPI
    −$1.90
    Revenue
    +171%

    Tidebreakers How it was measured

    Genre
    Mid-core strategy
    Geos
    US, KR, JP, DE
    Window
    16 weeks, D30 cohorts
    Measured with
    MMP + incrementality holdout + predictive LTV (D7 → D180)
    What failed
    Cinematic CGI trailers: high CTR, D7 ROAS 9%. Players expected the trailer, not the game.

    No. 003Tap to flip back

    See exactly how we did it →

  • Lucky Lanes

    Uncommon

    Social casino12 weeks

    Monthly spend
    $35K → $240K/month
    D30 ROAS
    41% → 63%
    CPI
    −$1.12
    Revenue
    +238%

    Lucky Lanes How it was measured

    Genre
    Social casino
    Geos
    US, CA, AU
    Window
    12 weeks, D30 payer cohorts
    Measured with
    MMP + value-based cohorts + geo holdout
    What failed
    Broad install campaigns: CPI −40%, payers −60%. Paused in week 3.

    No. 004Tap to flip back

    See exactly how we did it →

Swipe for more cards

Creative lab

Creatives are the new targeting.

Broad targeting hands the audience to the algorithm, and the ad decides who it finds. So we run creative like a lab: new concepts every week, clean tests against the control, budget only for winners.

Creatives tested this month
142
Winners
11
Win rate
7.7%

One creative, testedBeat the controlVideo: 36 tested, 3 winners

Video ad for Stackopolis, 12 seconds. It opens on the hook “Only 3% reach the clouds”, cuts to a montage of glossy blocks landing perfectly as the tower climbs, slows down for one near-miss where a slice breaks off, pulls back above the clouds and ends on an end card with an Install button.

StackopolisHybrid-casual

Gameplay montage

Real gameplay cut to the beat: a challenge hook, a run of perfect drops, one near-miss, then the end card. Twelve seconds, built to work with the sound off.

Hook rate
31%
IPM
8.4
CTR
1.9%
Status
Winner

36 video ads tested this month, 3 winners

Playable ad for Stackopolis. Tap or click the frame, or focus it and press Space or Enter, to drop each floor onto the tower. Ten floors or one miss ends the run on an end card with Install and Replay.

StackopolisHybrid-casual

Tap-to-stack playable

Go on, play it. Tap the frame to drop each floor and land it clean for a Perfect. Like every playable we ship, it ends on an install button.

Hook rate first tap
68%
IPM
14.6
CTR end card
4.8%
Status
Winner
  • Single HTML5 file, 2.1 MB, MRAID 3.0
  • Runs on AppLovin, Unity, ironSource and Mintegral
  • Tutorial hand until the first tap
  • End card at floor 10 or the first miss

12 playables tested this month, 2 winners

UGC ad for Kraken Kitchen, 10 seconds, shot selfie-style: a creator, back-lit so only a silhouette shows, holds a phone with the game to the camera while big captions appear word by word: “POV: you finally found a puzzle game that respects your time”. The camera pushes into the phone for real gameplay, then an end card with an Install button.

Kraken KitchenMerge puzzle

Creator POV

One creator, one take, captions burned in. It reads like a post, not an ad, so people watch it like a post.

Hook rate
38%
IPM
6.1
CTR
2.3%
Status
Scaling

Creator brief

  • The hook is the first word: POV
  • Phone in frame by 0:02
  • Footage from the real build, never a mock-up
  • No logo until the end card

24 UGC ads tested this month, 2 winners

Fake gameplay
Fake-gameplay ad for Tidebreakers, 11 seconds: a pin-pull puzzle under the hook “99% fail this”. A hand pulls the wrong pin and lava pours onto the knight, “Fail”; the clip rewinds, the hand pulls the other pin, gold pours down, “Level clear”, then an end card with an Install button.

TidebreakersMid-core strategy

Pin-pull: “99% fail this”

The classic fake-gameplay hook: show the fail, then the fix. Pin-pull isn't the core loop, so before we scaled it the studio shipped it as an in-game event. The ad shows what players actually get.

Hook rate
46%
IPM
12.4
CTR
3.4%
Status
Scaling
Judged on cohorts, not clicks
FormatIPMD7 ROAS
Real gameplay control7.918%
Pin-pull fake gameplay12.421%

20 gameplay and fake-gameplay ads tested this month, 1 winner

Lucky LanesSocial casino

Generated variants

  1. 64variants generated
  2. 16went live
  3. 1winner
4 hooks (rows) × 4 colorways × 2 symbol sets × 2 end cards (columns). Live Winner

One concept, 64 variants. A model pre-scores every one, the top 16 go live, and the market picks the winner. Same layout, so the test isolates the color and the words.

16 AI variants tested this month, 1 winner

  • Variant 05, Midnight colorway: navy and gold slot reels spin and stop on three sevens under the hook “Three 7s. One tap.”

    #05 Midnight

    Hook
    24%
    IPM
    5.2
    CTR
    1.6%
    Killed
  • Variant 22, Mint colorway: teal reels with bowling balls under the hook “Your lucky lane is open”.

    #22 Mint

    Hook
    27%
    IPM
    6.0
    CTR
    1.8%
    Testing
  • Variant 43, Coral colorway, the winner: coral and gold reels stop on three bowling pins, “Strike!”, under the hook “Strike it rich”.

    #43 Coral

    Hook
    35%
    IPM
    8.9
    CTR
    2.6%
    Winner
  • Variant 56, Cobalt colorway: blue and porcelain reels stop on three crowns under the hook “Roll for the jackpot”.

    #56 Cobalt

    Hook
    21%
    IPM
    4.4
    CTR
    1.3%
    Killed

StackopolisHybrid-casual

Same ad, three openings

Same body, three different first three seconds. Hook rate, the share of impressions still watching at 0:03, decides which opening gets the budget.

1.9× more viewers past 0:03. The challenge opener held 41% of them, the satisfying one 22%.

24 hooks tested this month, 1 winner

  1. Hook A, satisfying: a close-up, slow-motion perfect drop with a soft flash, captioned “Wait for it…”.

    A Satisfying

    “Wait for it…”

    Hook rate
    22%
    IPM
    5.9
    CTR
    1.4%
    Killed
  2. Hook B, fail: a leaning tower takes a bad drop and the top floors topple, captioned “Don't let it fall”.

    B Fail

    “Don't let it fall”

    Hook rate
    34%
    IPM
    7.6
    CTR
    1.8%
    Runner-up
  3. Hook C, challenge, the winner: the counter climbs from floor 97 to 99 high above the clouds while a block slides past, captioned “Only 1% reach floor 100”.

    C Challenge

    “Only 1% reach floor 100”

    Hook rate
    41%
    IPM
    9.8
    CTR
    2.2%
    Winner

Kraken KitchenMerge puzzle

One ad, four rounds

Winners are rarely found. They're iterated: one change per round, each version tested against the last winner.

3× the installs per impression. IPM went from 3.1 to 9.3 in four rounds.

10 iterations tested this month, 1 winner

  1. Version 1: a landscape gameplay capture letterboxed inside the vertical frame, no hook.

    v1IPM 3.1

    Baseline: landscape gameplay capture

    Hook
    14%
    CTR
    0.9%
    Retired
  2. Version 2: the same gameplay reframed so the merge board fills the vertical frame.

    v2IPM 4.6

    Reframed to full 9:16

    Hook
    21%
    CTR
    1.3%
    Retired
  3. Version 3: opens on a nearly full board with the hook “One move left?”, then one merge sets off a chain.

    v3IPM 6.8

    Hook in the first second: one move left

    Hook
    33%
    CTR
    1.9%
    Retired
  4. Version 4, the winner: a kraken tentacle reaches across the board to steal a dish under the hook “Save the kitchen!”, with an Install button.

    v4IPM 9.3

    Kraken fail-state opener, two-tap end card

    Hook
    39%
    CTR
    2.6%
    Winner

Every dollar. Every cohort. Every decision.

Stackopolis, a sample account. Change the range or the network: every number recalculates from the same cohorts.

Stackopolis UA dashboard

Hybrid-casual · iOS + Android · Worldwide

Sample data

MMP + network APIs · synced Oct 5, 06:00 UTC

Aug 6 – Sep 4, 2026Install cohorts matured to D30 · vs prev. 30D

Spend
$1.24M
up 26% vs $984K
Installs
685K
up 62% vs 423K
CPI
$1.81
down 22% vs $2.33
D7 ROAS
31.8%
up 4.4 pts vs 27.4%
D30 ROAS
74.2%
up 10.3 pts vs 63.9%
Revenue D30
$920K
up 46% vs $629K
LTV D30
$1.34
down 10% vs $1.49
Payback projected
Day 52
16 days sooner vs Day 68
74% recovered by D30

Daily spend & installs

By install date · All networks

Daily spend, installs, CPI and decisions by install date
Install dateSpendInstallsCPIDecision
Aug 6$34.0K16,058$2.12
Aug 7$34.5K15,876$2.17
Aug 8$36.8K16,055$2.29#96 retired: CTR down 40% in 3 weeks
Aug 9$37.9K16,725$2.27
Aug 10$35.1K16,714$2.10
Aug 11$35.6K17,402$2.05
Aug 12$37.6K18,417$2.04
Aug 13$36.1K17,826$2.03
Aug 14$37.6K18,205$2.06#121 paused: cheap installs, weak D7 ROAS
Aug 15$40.8K20,121$2.03
Aug 16$40.0K19,439$2.06
Aug 17$37.4K18,695$2.00
Aug 18$37.0K18,314$2.02
Aug 19$37.0K19,272$1.92
Aug 20$38.8K20,539$1.89#142 “Tower almost falls” launched
Aug 21$40.2K21,303$1.89
Aug 22$41.5K22,242$1.87#133 paused: IPM under 5
Aug 23$43.2K23,190$1.86
Aug 24$38.8K21,993$1.76
Aug 25$41.8K23,616$1.77
Aug 26$40.2K22,989$1.75
Aug 27$41.8K26,170$1.60#142 budget +35%: D7 ROAS above the 28% target
Aug 28$45.5K27,679$1.65
Aug 29$49.9K29,656$1.68
Aug 30$51.5K32,046$1.61
Aug 31$46.7K30,508$1.53
Sep 1$48.5K31,813$1.52#151 playable moved to Scale; #155 iteration live
Sep 2$50.5K32,788$1.54
Sep 3$52.1K35,146$1.48
Sep 4$51.5K34,214$1.50

Spend by network

Every split adds up to the total

Spend, share, installs, CPI and ROAS by network
NetworkSpendShareInstallsCPID7 ROASD30 ROAS
AppLovin$412K33%178K$2.3135.9%83.2%
Meta$325K26%158K$2.0633.3%78.5%
Google$228K18%140K$1.6329.4%68.9%
TikTok$150K12%131K$1.1524.2%56.7%
Unity$125K10%78.7K$1.5928.0%63.9%
Total$1.24M100%685K$1.8131.8%74.2%

Cohort ROAS ladder

Revenue ÷ spend by day since install

Cohort ROAS by day since install, against target
Day since installCohort ROASTarget
D06.1%5.0%
D111.4%9.0%
D319.6%17.0%
D731.8%28.0%
D3074.2%66.0%
Day 52 (projected)100% (payback)Payback by day 60

Retention

Share of installs still playing on day N

Retention by day since install
DayStackopolisGenre median
D141.0%34.0%
D327.4%19.6%
D718.0%12.0%
D1411.9%7.7%
D307.0%4.5%

Creative performance

Top creatives by spend in this view · status as of Oct 5

Creative performance: format, spend, CPI, IPM, hook rate, D7 ROAS and status. Sortable by spend, CPI, IPM, hook rate and D7 ROAS.
CreativeFormatStatus
#142Tower almost fallsGameplay videoGameplay video$180K$1.389.438%41.0%Scale+35% budget
#138Perfect streak ×10Gameplay videoGameplay video$152K$1.768.231%33.4%ScaleHolding budget
#129Day 80 of my cityUGC videoUGC video$68.7K$1.387.135%23.3%TestNew hook in test
#151Stack 5 floorsPlayablePlayable$47.7K$1.5014.6—38.8%Scale+20% budget
#147City from aboveAI videoAI video$34.7K$1.497.227%30.2%TestBudget capped
#121Skyscraper failFake gameplayFake gameplay$27.6K$1.2513.339%12.1%KillPaused Aug 14
#133Satisfying slicesStaticStatic$17.7K$2.564.4—19.4%KillPaused Aug 22
#155Wrong block!Gameplay videoGameplay video$7.8K$1.329.043%35.2%TestIteration of #142
Other creatives134 more in this viewMixed$704K$2.147.625%30.6%
Total 142 creatives$1.24M$1.818.230%31.8%

Hook rate = 3-second views ÷ impressions on Meta, TikTok and YouTube. IPM = installs per 1,000 impressions. The bar marks D7 ROAS against the 28% target.

Sample data: a fictional game with illustrative numbers.

Spend and impressions come from the network APIs; installs, revenue and retention from MMP raw data. Payback (day 52) is a projection.

No black-box reporting.

Your MMP, your ad accounts, your raw data — the same numbers we look at.

We don't work like an agency.

You get a squad:

  • Growth Lead

    In your squad

  • UA Manager

    In your squad

  • Creative Strategist

    In your squad

  • Creative Team

    In your squad

  • Data Analyst

    In your squad

  • Instead of: an account manager

    Not included

Shop

Prices in public. Rare in this niche, which is the point.

Best for launches

Startup pack

Bundle contents

  • Account setupAd accounts and tracking set up from day one. You own them.
  • Creative productionThe launch set of ad creatives.
  • UASetup, infrastructure and ongoing management.
  • ASOSetup and infrastructure for the store listing.
  • SMMSocial media set up for launch.

From $12K per month

Get the startup pack

  • Creative sprint

    30+ concepts, 120+ variants. Playables quoted per build.

    from$14Kper month

  • ASO & CPP program

    Keywords, screenshots, CPPs, product page tests, 30+ locales.

    from$6Kper month

  • Managed UA

    $50K a month media minimum. You own the accounts.

    10%of spend

  • Full funnel

    UA, creative, ASO and measurement on one P&L.

    Customscoped together

Your game has data. Let's turn it into scale.

No sales pitch. Bring us your KPIs.