I am building a side project and the question of how to charge for it has been on my mind longer than I expected. Not because the options are unclear, but because the sequencing is not obvious. Getting the revenue model wrong at the wrong time seems to be one of the more common ways apps stall, and I wanted to understand the shape of that failure before I ran into it.
Here is what I have worked out so far.
What kind of product determines which model fits?
The usage pattern is the first filter.
If a product has a short, intensive lifecycle (someone uses it for a few weeks or months, then leaves), recurring subscriptions are a hard sell. The user already knows they will churn. One-time purchases or pay-per-use match the psychology better because the value is bounded and the pricing reflects that.
If a product earns habitual, ongoing use, subscriptions make sense. The user's value accumulates over time and the pricing reflects that. The risk is churn from users who subscribed before forming the habit and never came back.
If the product targets businesses rather than individuals, the ceiling on pricing shifts dramatically. A $200/month tool that saves a team five hours a week is cheap. The same tool sold to a consumer is not. Business buyers do not spend personal income.
I have been reading through case studies of failed freemium products and the pattern that shows up most often is a mismatch between how users actually experienced the product and the pricing model the founder assumed would fit. Subscription models applied to utilities people use once. One-time pricing on products where the value only compounds after months of use.
What are the actual options?
Advertising
Display ads are the lowest-friction path to early revenue. No auth, no payment infrastructure, no architecture change. You add a script and start earning.
The ceiling is lower than most founders expect going in. Meaningful ad revenue requires sustained traffic, and the CPM depends heavily on geography and niche. Education and finance tend to earn more per visit than general content. Consumer apps with traffic spread across many geographies often see underwhelming rates.
Ads also carry a real UX cost. Users tolerate them but they create friction that affects retention. Placement matters: ads between sessions or on hub pages damage the experience less than ads inside the core workflow.
Native or sponsored placements from brands aligned with your niche tend to outperform generic display in both revenue and user experience. A relevant recommendation looks like a service. A banner looks like a banner.
Freemium
Keep the core product free, gate premium features behind a paywall.
The trap I kept reading about is gating the wrong things. If you gate features users need to get value from the product, they leave before they form an opinion worth paying to extend. If you gate features users want after they have already gotten value, conversion is easier because the user already believes the product works. The free tier has to be genuinely useful. The premium tier should solve something the user has already bumped into as a real problem.
Pricing for freemium tends to land where the monthly cost does not require approval but is still meaningful. Annual pricing at a discount converts users who have already decided and reduces churn.
One-time or lifetime purchase
A flat fee for lifetime access is psychologically cleaner for users with bounded needs. They pay once, there is no subscription anxiety, and the transaction is complete.
This seems to work best for tools and utilities where the use case has a defined end. It also converts well in markets where recurring billing feels uncomfortable or where users have been burned by subscription creep elsewhere.
The trade-off is revenue predictability. One-time sales peak around launches but do not produce a stable monthly baseline the way subscriptions do.
Pay-per-use
Charge for specific actions rather than access to the product overall.
This fits when the value of each transaction is clear and bounded. AI-generated feedback, document exports, detailed analysis, a specialized report. The user pays for the output, not for the seat.
It is also a good fit for premium features that not every user will want regularly. It lowers the barrier to trying the feature and lets volume users self-select before you offer them a bundle or upgrade path.
Digital products
PDFs, templates, resource packs, and other downloads sold outside the main app. Platforms like Gumroad or Lemon Squeezy handle the transaction. No payment infrastructure required on the product itself.
The appeal is low lift for early cash. The ceiling depends on how well the asset is positioned and whether you already have an audience to sell into.
Affiliate revenue
Earn a commission when your users convert on a product you recommend.
This works when the recommendation is a natural next step for the user. If you are building a tool for a specific preparation task, recommending the books or courses your users would want anyway reads as a service, not a pitch. Conversion is higher and the trust cost is lower.
Disclosure is not optional. Users who discover undisclosed affiliate arrangements lose trust faster than they would have if the arrangement had just been stated upfront.
Donations and tips
Ko-fi, Buy Me a Coffee, and similar platforms let users support a product without a formal subscription.
From everything I have read, donations work as supplementary income, not as a strategy. They are a floor, not a ceiling, unless you have built an unusually loyal audience that treats support as part of the relationship.
What order does this happen in?
Building a full subscription system from the start, before you know what users value, seems like a common mistake. It locks you into a pricing model before you have the signal to know if it fits, and it adds complexity before you have the traffic to justify the lift.
The sequencing that keeps coming up in the resources I have been reading:
- Start with something that requires no engineering. Donations, affiliates, a recommended resources page. These tell you whether users trust the product enough to transact at all.
- Add a digital product if you have something worth packaging. A useful download tests price sensitivity without requiring payment infrastructure.
- Add pay-per-use or AI-powered premium features once you understand which actions users actually want to pay for.
- Build full subscription infrastructure when you have enough traffic, retention data, and conversion signal to make the investment sensible.
Ads belong late for most products. The revenue is too thin to justify the UX cost until the traffic numbers support it.
Where does willingness to pay actually live?
The pattern I keep coming back to: for a product with a functional free tier, the strongest monetization lever is a paid feature that answers a question the free version raises but cannot answer. The free tier creates belief that the product works. The paid tier answers "but is my specific answer good enough" or "what should I do next" at a resolution the free version cannot reach.
That gap is where willingness to pay concentrates. Not in access, but in certainty.
Everything else, ads, affiliates, downloads, is real but secondary. If a user has a question only the paid version can answer, the conversion argument is already present in the product. You are not selling anything. You are completing something.