Jon Yongfook: From 20 Failed Products to a $1M ARR API

Introduction

The boldest product decision Jon Yongfook ever made was an act of deletion. In 2019, the Singapore indie hacker had a working tool called PreviewMojo — it auto-generated social media preview images, had hit #2 Product of the Day on Product Hunt, and was making about $400 a month. Ramen money. Then he noticed a pattern in the user requests: "Can it make Instagram posts? Pinterest pins? Ad banners?" Marketers were spending hours hand-crafting every visual, and they wanted the tool to do more of it.

So Yongfook killed the user interface. He deleted the dashboard — the thing every SaaS founder is told is the product — and turned the whole thing into a pure REST API that plugged into Zapier, Airtable, and any other workflow tool. Within seven days, a $49 subscription pinged his Stripe account. That was the birth of Bannerbear, an API that auto-generates images and videos for marketing teams. Bootstrapped, solo-founded, it crossed $1 million in annual recurring revenue — built on the wreckage of more than twenty products that failed.

The Story

Before Bannerbear, there was a decade of almost nothing. Yongfook had spent 10+ years working for multinationals and startups across Asia, then quit the corporate career to build his own things. The things did not work. By his own account on the Ahrefs podcast, he went through 20 to 30 failed projects before finding product-market fit — a graveyard most founders would rather not itemize.

In mid-2018, stuck in that graveyard, he did something that looked like surrender and turned out to be strategy: he signed up for the Indie Hackers "12 startups in 12 months" challenge. The goal was not to build twelve unicorns. It was to regain the ability to ship fast and find out what works. Out came a blur of small launches: Beatrix, a virtual social media assistant; Zipsell, an open-source platform for selling digital downloads; Promomatic, a tool for generating App Store design assets; Montage, a design component library; Tech Jobs Asia, a niche job board; Hyperjump, a form builder for the financial industry; Tropic, a collaboration app for teams across time zones. None of them became Bannerbear. All of them rebuilt the shipping muscle.

PreviewMojo was one of those twelve — the one that refused to die quietly. It was almost embarrassingly narrow: paste a URL, get a social media preview image. It launched on Product Hunt, took #2 Product of the Day, and settled into roughly $400 in monthly recurring revenue. Yongfook could have spent a year polishing its dashboard, adding templates, running ads. Instead he listened to what users were actually asking for, and the requests all pointed past the interface: they didn't want a better preview-image app, they wanted a machine that could generate every visual their marketing needed, inside the tools they already used.

So he did the thing that feels insane until you see the numbers. He removed the interface entirely and exposed the engine as an API. Marketers could now generate images and videos without writing a single line of code — or without ever opening his product — by wiring Bannerbear into Zapier and Airtable. The first $49 subscription arrived within a week of the pivot. The lesson he later distilled from it is worth quoting in full: when users pull you somewhere, sprint.

What followed was not a viral explosion but a method. Yongfook's most influential contribution to indie-hacker culture may be the framework he calls the 50/50 split: alternate coding weeks and marketing weeks, so building is never prioritized over distribution. Week one: ship code. Week two: write content, post on X, build free tools. Within two years of adopting that rhythm, his revenue climbed from a few hundred dollars a month to $10,000, then past $50,000 a month. Bannerbear's open-startups page documented the climb in public — milestones from $10K to $50K MRR on bannerbear.com/open — and third-party trackers later showed the business well past the $1M ARR mark, still bootstrapped.

The 50/50 split forced a pricing reckoning early. Yongfook's first instinct was the classic indie mistake: price Bannerbear at $9 a month. At that price, meaningful revenue would have required around 2,000 customers — and at typical internet conversion rates near 1%, that meant finding roughly 200,000 free users first. Insanely tough. So he priced up instead: plans from $49 to $399 a month. At one public snapshot, 284 subscribers added up to $20,000 in monthly recurring revenue. Fewer customers, more revenue per customer, and support tickets he could actually answer himself.

His marketing weeks were as systematic as his coding weeks. He built free tools as lead magnets — small, genuinely useful utilities that ranked on Google and funneled users toward the paid API. He wrote "versus" and "alternatives" comparison pages, the unglamorous SEO work that captures buyers at the exact moment they're choosing. And he posted spicy takes on X to keep the top of the funnel warm. None of it required a marketing team. All of it compounded.

The empire stayed quiet and kept growing. Bannerbear's success funded horizontal expansion into adjacent products — ClipCat, a video-rendering engine for short-form content, and RoboRabbit, a data extraction tool — each one sharing the same API-driven automation DNA. Yongfook also wrote the Growth Hacking Handbook, became a regular speaker at tech conferences across Asia, and kept the operation lean enough to run from coworking spaces across Southeast Asia and Europe. No venture capital, no growth-at-all-costs phase, no pivot-to-enterprise theater. Just a founder who spent years learning what not to build, then built the one thing users were already asking for.

Lessons

Jon Yongfook's run offers three mechanics worth stealing: 1. Let the users pull you — then sprint: PreviewMojo was a $400/month curiosity until user requests revealed the bigger job: auto-generate every marketing visual, inside existing workflows. Copy it: add a channel where users tell you what else they'd use your product for, and treat the pattern in their requests as a roadmap written by the market. Don't polish the interface — follow the pull, even if it means deleting what you built. 2. Run the 50/50 split: one week shipping code, one week shipping distribution. Most indie hackers out-build their marketing for months, then wonder why nobody comes. Copy it: block your calendar into alternating weeks now. A marketing week means content, free tools, comparison pages, and public build logs — not "thinking about growth." Yongfook went from hundreds to tens of thousands in monthly revenue on this rhythm alone. 3. Price up, customer count down: at $9/month he'd have needed ~2,000 customers (and ~200,000 free users to find them); at $49–$399 he needed ~300. Copy it: divide your revenue target by your price. If the customer count comes out above 500, you don't have a marketing problem — you have a pricing problem. Raise the price until the number fits a founder who also does support.

Closing

Jon Yongfook spent the better part of a decade building products nobody wanted, then built the one product users were asking for — and had the nerve to delete its interface. The API was never the insight; listening was. Twenty failures taught him to ship fast, one pivot taught him to follow the pull, and a simple weekly rhythm turned a $400 side project into a $1M-a-year business. The market writes the roadmap. The founder's job is to ship fast enough to read it.