Loading...
Most founders reach for subscriptions first. Build a product, charge every month, and track monthly recurring revenue.
Sometimes that is the right choice. Sometimes it makes customers pay every month for a tool they only need now and then. This post looks at four ways to charge and shows how a few founders made money from much stranger ideas.
Subscriptions give a business more predictable revenue. They also make sense when customers keep getting value from the product.
Think about hosting, email software, or a tool your team opens every day. The service keeps running, so a regular payment feels fair.
But many products are not used that way. A founder may research an idea for a week and then stop for a month. A designer may need a file converter twice a year. Charging those customers every month gives them a reason to cancel.
The question is not, "Can this be a subscription?" Almost anything can. Ask, "Does the customer get useful value every month?"
There is good evidence that people review recurring charges and switch services when the value drops.
In Deloitte's 2025 US media survey, 39% of consumers said they had canceled at least one paid streaming service in the previous six months. Another 24% had canceled and then joined the same service again. A $5 price increase would make 60% likely to cancel even their favorite service.
A separate Bango survey of 5,000 US subscribers found that the average respondent paid for 5.4 subscriptions. The full report says 36% often pause and restart services.
These studies focus mostly on consumer subscriptions and streaming, not small business software. They do not prove that every SaaS customer behaves the same way. They show something simpler: a recurring payment does not guarantee recurring use.
That matters when people use your product for short projects. In that case, credits or a one-time payment may fit better.
Cost is not the only pressure on the old models. AI is now reshaping what a "user" even means, and that hits per-seat pricing hardest.
In September 2026, Basecamp scrapped per-user pricing. Founder Jason Fried called it "downright antique" in the agentic age. The new plans start at $59 a month and include unlimited users and agents. The cap is on projects now, not people.
His reason was simple. You can now bring AI agents into Basecamp as users, through its command-line tool today and MCP soon. Charging for each agent felt as wrong to him as charging for each person.
This is one company's call, not a rule for everyone. But it points to a real shift. When software does more of the work on its own, a price per seat measures the wrong thing.
Customers buy a balance or pay for what they use. They spend money when they need the product and pay nothing when they do not.
This is how PreVibe works. People often research an idea heavily for a few days, then return when they have another idea. Their unused credits stay in the account, so there is no subscription to cancel.
Credits work well when:
The tradeoff is less predictable revenue. Customers also need to understand what one credit buys, so keep the system simple.
The customer pays once for a product, download, or license. There is no renewal to manage.
This can bring in cash quickly and works well for templates, courses, downloads, and software with low running costs. It is harder when every customer creates years of hosting or support work. Be clear about whether the price includes future updates and how long you will provide support.
The product stays free for users, while advertisers pay to reach them. This only works when you have an audience that sponsors want.
Pieter Levels used this model for his free browser flight simulator. It sells ads inside the game and also offers a $29.99 fighter plane. Levels says the game reached $87,000 in monthly revenue 17 days after launch, which he described as a $1 million annual run rate.
That does not mean the game earned $1 million in 17 days. It means one month's revenue, multiplied by 12, was running at about $1 million a year. The number is also reported by the founder, not from audited accounts.
The same founder made a bolder move in September 2026. He made Nomads.com, his community for remote workers, free to join after 12 years and 43,252 paid members. He kept a $1 charge only to block spam. Any real price, he said, caps how fast a community can grow, so he would rather grow the audience and let sponsors pay for reach. He also believes communities are harder for AI to copy than software, and he wants this one as large as possible.
The useful lesson is the same in both cases: if a free product attracts the right people, you may be able to charge sponsors instead of users.
Subscriptions are still a good option when you provide steady, ongoing value. They are especially useful when you also have regular costs such as hosting, data, support, or new content.
Do not choose them only because monthly recurring revenue looks good on a dashboard. Choose them when a customer can look back each month and see what they paid for.
| Model | How customers pay | Good fit | Main drawback |
|---|---|---|---|
| Subscription | A fixed amount on a schedule | Products used throughout the month | Customers may cancel during quiet periods |
| Credits or usage | For a balance or actual use | Work that happens in bursts | Revenue can vary from month to month |
| One-time payment | Once, up front | Downloads and products with low running costs | You must keep finding new buyers |
| Ads or sponsorships | Advertisers pay for attention | Free products with a clear audience | You need enough reach to interest sponsors |
You can also combine these models. For example, a product might offer a subscription to regular users and credit packs to occasional users.
These ideas are not plans to copy. They show that people will pay for attention in surprising places.
In August 2026, German developer Jonathan Wilke launched outbid.lol, a public leaderboard where a company's position depends on how much it pays. Wilke says he built the first version in about three hours.
The site's launch announcement reported more than one million visitors and $120,000 in bids during its first 48 hours. The top position cost $13,005 at that point.
Those figures came from the founder and the site's own data. They are gross payments, not audited profit. Even with that caveat, the launch shows how the bidding itself became part of the story and brought more people to the page.
Sergey Nazarov took the leaderboard idea in a stranger direction with sponsor.baby. Companies pay for a place on a board that shows the current "baby name." Nazarov launched it with his own product, ScreenCharm, in first place and credited Wilke for the idea.
The name is a joke, not a real naming contract. The site's rules say a payment buys a leaderboard listing and does not legally require Nazarov to give a future child the winning name.
That distinction matters. The product sells a funny public story and a link, not actual naming rights.
French founder Dagobert Renouf sold ad space on the suit he wore at his wedding. Twenty-six startups paid to appear on it. Inside placements started at $100, while visible spots on the outside cost $300 to $2,000.
Renouf told People that he sold $10,000 in sponsorships. After the suit and taxes, he had about $2,000 left. One sponsor was so impressed by the campaign that it hired him.
He did not have a software product to sell in this case. He had an event, an online audience, and a story that sponsors wanted to join.
Start with how and when customers get value from your product.
Then check the costs. A one-time price is risky if you must host a customer's data for years. Usage pricing is risky if heavy use costs you more than you charge. Ads are a poor fit if they make the product less useful or hurt trust.
You do not have to choose one model forever. Start with the simplest one you can explain, watch how customers use the product, and change it when the evidence tells you to.
Before choosing a model, look at how similar products charge. Their pricing will show you what buyers already understand and where there may be room for a different offer.
PreVibe finds more than 20 products related to your idea and shows their pricing, revenue and traction signals, and growth tactics from live sources. It is built for solo founders and small teams, and your first research is free.
A subscription is useful when customers get value every month. It is frustrating when they do not.
Use credits for work that comes in bursts, one-time payments for products with low running costs, and ads when sponsors want access to your audience. Pick the model that matches how people use the product, then check it against products already in the market.
For more on funding a small business, read our guide to bootstrapping and venture funding.
Bootstrapping means growing a startup on revenue and your own savings instead of investor money. Here is what it is, and why it often beats the venture path.
Your ideal customer profile is the single buyer your product is built for. Here is why a fuzzy ICP quietly kills good ideas, and how to find a sharper one.
Follow the proven sequence of research methods successful founders use to go from idea to product-market fit.