$488 a Month for a Year, Then $10,455: Jon Yongfook Published Every Step

Jon Yongfook left his job in January 2019. By September of that year his first product was earning $400 a month. In December it was $472. In February 2020, five months later, it was $488. Anyone reading a dashboard like that would reasonably conclude the thing was dead.

By January 2021 it was $10,455 a month. By a post he published in July 2023, Bannerbear was at $50,000 a month, still a small team, still built around a single narrow idea: an API that generates images and videos automatically, so other people's software can make graphics without a designer. What makes it useful as a case study is that Yongfook wrote the whole thing down as it happened, including the flat 18 months.

The flat part is the story

In his own account of getting to $10k a month, the monthly figures read like this: $400 in September 2019, $472 in December, $488 in February 2020, $6,109 in October 2020, $10,455 in January 2021.

Look at the gap between February and October 2020. That is where the business was actually made, and nothing in the first three numbers predicts it. Two decisions closed the gap.

The first was scope. The original product, Previewmojo, automatically generated social preview images for a website's pages. Useful, narrow, and priced for individuals. In March 2020 he rebranded to Bannerbear and released a REST API, turning a finished tool into a building block other developers could wire into their own products. In the same month he pulled out of the Shopify app store, where the product had not found traction.

The second was price. Yongfook is blunt that starting at $9 a month was a mistake. Today the plans are $49, $149 and $299 a month. Same core idea, an order of magnitude difference in who the customer is. Builtplain has made the argument in the abstract in setting a first price you can defend; Bannerbear is the version with dates attached.

One week coding, one week marketing

The operating rhythm he describes in the $50k MRR post is a two-week cycle: one week writing code, one week doing marketing, repeated indefinitely. Blog posts, tweets, forum answers, the newsletter.

This is duller than any growth tactic and considerably more effective than most, because it removes the decision. A solo founder who "does marketing when there is time" does no marketing, since there is never time. A solo founder on a fixed alternation ships roughly 26 weeks of marketing a year whether or not the mood strikes.

Two specific channels came out of those marketing weeks. Documentation and tutorials were treated as a conversion tool rather than an afterthought: "The more tutorials I have, the more users convert as they understand the product better." And free micro-tools, a certificate generator, a Tweetagram tool, were built to pull in search traffic and links from people who would never pay but who share things.

The first hire tells the same story. In May 2021 he brought in a freelance writer for tutorial content. Not a salesperson, not a developer. The thing that converted got the money.

The second climb was faster than the first

Zero to $10,000 a month took about a year of active effort after launch. Ten to twenty thousand took roughly six months, from February to July 2021, as he recorded at the time.

Two shifts drove it, both mundane. He started building what paying customers asked for instead of what he found interesting, and he deliberately went after higher-tier customers because they churned less. An API business at $49 has a support burden per dollar that an API business at $299 does not.

Integrations did the rest of the work. Integromat arrived in March 2021, and Zapier and Make connections put Bannerbear in front of the no-code community, where people who cannot write code but need automated images go looking. Starter Story's breakdown of the business notes the same effect: the integrations produced word of mouth and free tutorials written by other people.

Being a component inside somebody else's workflow is an underrated distribution strategy for a technical product. You do not have to be the app anyone opens. You have to be the box they drag into a diagram.

What this cost, and what it did not

Bannerbear was not a hobby project that got lucky. Yongfook had 20 years as a designer and programmer before starting, and he had already spent about a year on a rapid-fire series of product launches before landing on this one. The idea was not divine inspiration either: it came from a problem he understood, in a category he had worked in.

The failures are in the record too. The Shopify app that did not work. The $9 pricing. A DDoS attack that forced an infrastructure migration he now says should have happened earlier. The early API launch that annoyed some existing customers.

His stated rule on hiring is worth copying verbatim: hire "only when you are feeling the pain". Not when a plan says a role should exist, but when a specific piece of work is visibly hurting.

The checklist a reader could actually run

1. Sell a component, not just an app. An API or integration lets other people's products carry yours.
2. Price for businesses. A $9 plan attracts users who cost more in support than they pay.
3. Fix a marketing cadence you cannot argue with. One week on, one week off, forever.
4. Treat documentation as sales collateral. Every tutorial is a page that answers "can this do my specific job?"
5. Build small free tools that solve one instance of the problem your paid product solves generally.
6. Get into the integration directories your customers already live in.
7. Publish your numbers. His milestone posts were themselves a marketing channel, and they made him findable by every founder searching for the same milestone.

What is not on offer here

Two caveats. Bannerbear's open metrics page now reads "Bannerbear is was an open startup with public metrics, we now only share selected metrics", so the published figures stop at the milestones above. Nothing in this article should be read as a statement about what the business earns today.

And the timeline is the real lesson. From leaving a job in January 2019 to $10,000 a month in January 2021 is two years, with a long stretch in the middle where the revenue chart looked like a flat line and the correct response was to change the product rather than quit. That is the part of the one-person software company that the milestone screenshots leave out, and Yongfook is one of the few founders who bothered to write it down while it was happening.