I have read a lot of "subscriptions are eating mobile" pieces and a lot of "subscriptions are stalling out" pieces and a lot of "hybrid monetization is the future" pieces. They are all sort of right and sort of wrong in the same way.
The framing is the problem. Subscription versus ads is not a single industry-wide debate to be won. It is a context-specific question that depends on the app category, the user behavior pattern, the LTV math, and the operational complexity each model can support. The same answer does not apply across categories and arguing as if it should is what makes most of the writing on this topic so unsatisfying.
Subscriptions work where session frequency justifies recurring value
The apps where subscriptions work are the apps where users want to use the app frequently enough that recurring value is real. Productivity apps, learning apps, fitness apps, content apps with regular updates. The user perceives ongoing value because they actually use the thing on an ongoing basis.
The math works because LTV scales with retention. A user who stays subscribed for two years is worth twenty-four times a single month's price minus the customer acquisition cost. The leverage is real if the retention is real.
The risk is that the retention often is not real. Most subscription apps see significant churn in the first three months. The LTV models that justify high CAC are usually overweighting the long tail of retained users in ways that don't survive contact with reality.
Ads work where session count is high but engagement is shallow
Ad-supported models work where users open the app a lot but do not have a strong sense of recurring value worth paying for. Casual games, utility apps, social-adjacent apps, content browsers.
The economics depend on session count and ad inventory rather than retention price. A user who opens the app five times a day and sees three ads per session generates revenue regardless of whether they would have paid for a subscription.
The downside is that ad ARPU is much lower per user than subscription ARPU and is highly dependent on ad market conditions. A category that monetizes well at $10 eCPM falls apart at $5. The volatility is real and is mostly outside operator control.
Hybrid models work for specific reasons, not generally
The hybrid pitch — show ads to free users, offer subscription to remove ads — sounds great in the deck. In practice it works in some categories and not others. The pattern matters.
It works when the ad experience is genuinely annoying and the subscription removes a real friction. Music apps did this well. Video apps did this well. The free experience was bearable; the paid experience was meaningfully better.
It works less well when the free experience is fine and the subscription does not add enough value. Many casual game hybrid models are in this bucket. The free game is fun. The subscription removes some ads. Most users do not perceive the value differential as significant enough to convert at meaningful rates.
The hybrid approach also adds operational complexity. You need to manage two pricing models, two LTV calculations, two retention curves, and two churn dynamics. Smaller teams underestimate this overhead and end up doing both poorly.
The actual question is about your specific app
The question that produces useful answers is not "subscription or ads" but "what does my user behavior pattern actually support and what can my team operationally execute."
Look at session count and session length distribution. Look at engagement depth on key features. Look at the retention curve at 7, 30, and 90 days. Look at what users are doing in the app, not what you wish they were doing.
If you have a high-frequency app with shallow engagement, ad monetization is probably your default. Subscriptions might layer on for a small power-user segment but will not be the primary revenue driver.
If you have a lower-frequency app with deep engagement, subscriptions are probably your default. Ads might add some marginal revenue but will dilute the experience and probably hurt retention.
If you genuinely have both — high frequency and deep engagement — you can run hybrid, but you should be honest that this is rare and you should have the operational capacity to manage two revenue lines.
What changes with iOS subscription policy
Apple's subscription terms continue to evolve. The 30 percent first-year cut, the 15 percent second-year cut, the various carve-outs for small developers. The math has shifted multiple times and the next shift is probably coming.
For teams running subscriptions, the platform fee is a significant part of the LTV calculation. It is also outside your control. The teams that have built their LTV models with a pessimistic view of platform fees have generally been less surprised than teams that built models assuming current rates would persist.
The same applies on Android with Google Play subscriptions. The platform-level economics are a meaningful constraint that teams sometimes underweight when modeling subscription business cases.
What the pattern of recent debates actually shows
The recurring nature of the subscription-vs-ads debate suggests that neither model has converged on dominance. That itself is informative. If subscriptions were clearly winning, the ads debate would have been settled by now. If ads were clearly winning, the subscription debate would have been settled.
The reality is that both models work in different contexts and the success or failure of any individual app depends much more on category fit and execution than on which monetization model the team picked.
The teams that are doing well in mobile monetization are the teams that picked the model that fits their specific context, executed it competently, and stopped agonizing over the meta-question of which model is "winning." The teams that are doing badly are the teams that picked based on industry trend pieces rather than based on their actual app dynamics.
The most useful question to ask about your monetization is not "what is the industry doing" but "what does my specific user behavior pattern support and what can my team execute." That question gets useful answers. The industry-trend question rarely does.
