Subscription, One-Time, or Ads — Which Model Actually Fits Your Mini-Game?
People pick a monetization model the way they pick a logo — early, on vibes, and almost never again. Then they spend six months optimizing the wrong thing. The model is not a setting you toggle; it decides who your game is for, what you are allowed to build, and how much of your revenue you will hand back. Before you wire up a paywall, read the shape of the money, not the headline number.
Q: Why is "the model with the highest LTV" usually the wrong answer?
Because LTV is an average, and averages hide the two things that actually kill you: timing and variance. A subscription with a $30 predicted LTV looks better than IAP with a $12 LTV — until you realize the subscription's money arrives over eleven months, your churn model is a guess, and 40% of that LTV is a number you will never collect because the player cancels in week three. Meanwhile the IAP game got its $12 in minute four, and you can reinvest it into the next title. Paper LTV also quietly assumes you keep the money. Subscriptions carry the highest refund and chargeback exposure of the three, so the gap between "modeled LTV" and "cash in the bank" is widest exactly where the headline looked best. Pick the model whose revenue shape matches the game you actually built, not the spreadsheet you wish were true.
Q: What does the revenue "shape" of each model actually look like?
Subscription is a thin recurring line that decays as churn compounds. Its strength is predictable cash and a reason to keep shipping. Its weakness is that you have promised the player a service, so every content gap reads as a broken promise. One-time or premium is a single spike at install — clean, no churn math, but you have bet the whole game on the store listing and you have no built-in reason to return. Ads and rewarded placements are a slow metronome: many tiny payments from many players, heavily concentrated in the minority who watch a lot, and relentlessly sensitive to fill rate and eCPM. Hybrid — IAP plus rewarded ads — is what most survival games actually run, but it demands you design the ad as a mechanic, not a tax, or you train players to hate the button.
Q: How is refund and fraud risk distributed across the three?
This is where the models diverge more than the LTV charts admit. Subscription has the worst clawback tail: a player disputes month four's charge and the processor pulls it back from your balance, sometimes months after you have spent it. One-time purchases are cleaner but still face "item not received" disputes you must answer with delivery proof. Ads carry almost no direct refund risk — you are not charging the player — but they export the risk sideways: ad fraud, spoofed impressions, and the platform's own policy strikes that can zero your account. The honest framing: subscription and IAP make you a merchant with a chargeback ledger; ads make you a publisher with a fraud problem. They are different risk desks, and most small teams staff neither.
Q: How do you actually decide which fits your game?
Stop with the category and look at your retention curve and session shape. If day-30 retention is under 8% and sessions are two minutes, subscription will starve — you are asking players to pay for a relationship that ends in a week. That game wants IAP for cosmetics or a tight premium price. If you have a long-lived meta and weekly content, subscription starts to make sense, but only if your content cadence can survive the promise. If your strength is huge top-of-funnel volume with thin per-user depth, ads or rewarded-IAP is where the money is, and your job becomes fill rate, not conversion. The table below is the decision wedge most teams skip.
| If your game has… | The model that fits | Why | The trap to avoid |
|---|---|---|---|
| <8% D30, 2-min sessions | IAP / premium | You monetize the install, not a relationship | Building a subscription on a churn cliff |
| Long meta, weekly drops | Subscription | Recurring pay for recurring value | Letting content cadence slip — cancels follow |
| Huge volume, thin depth | Ads + rewarded IAP | Many small pays beat few big ones | Treating ads as a tax, not a mechanic |
| Mid everything | Hybrid | Diversifies the risk desk | Two half-built systems, neither tuned |
The model is the first line of your P&L, not the last. Pick it from the retention curve you measured, not the one you pitched.
