Most course creators on Udemy obsess over the same thing: how do I get more students to buy my course?
Rob Percival wasn’t thinking about backend systems when he launched. He was a math teacher with a web hosting side business trying to make $3,000 a month.
But something unexpected happened after his students bought, and that accident turned a single coding course into over $5 million in revenue.
The numbers look impressive on the surface. But the real story isn’t about course sales. It’s about what he built behind them – a recurring revenue backend that kept paying long after each course sale ended.
Here’s what he built, and what you can actually use.
The Hidden Asset He Already Owned
When Rob Percival left his job as a math teacher in Cambridge in 2012, he wasn’t chasing course revenue. He was running a small web hosting company called Eco Web Hosting that he’d launched as a side project back in 2006.
Nothing flashy, just enough monthly recurring revenue to replace his teaching salary.
“It was my first business that kind of grew organically without me having to do a great deal of marketing,” he said. “It turned out that was something that a lot of people wanted.”
For the next two years, he juggled freelance work and app ideas while the hosting company paid the bills. Nothing else clicked.
But that hosting business was building something he hadn’t yet recognizd: eight years of customer relationships with small business owners. People who worked with websites every day but had never learned to build them properly.
That database would become the launchpad for everything that followed.
In the summer of 2014, Rob partnered with a friend named Ben Tristem to run an in-person coding bootcamp for teenagers. Getting students through the door was hard work for modest returns.
One afternoon, Rob walked into Ben’s office and noticed Udemy on the screen, an online marketplace for digital courses. Ben was already teaching on the platform with moderate success.
Rob did what most people do when they discover Udemy. He looked at the top-selling courses and ran the numbers on what instructors were earning. The figures were significant.
Udemy already had millions of students browsing for courses. The top instructors were earning six figures. And the existing web development courses weren’t very good.
They were dry and technical. Focused on explaining languages rather than building things.
Rob had spent ten years teaching teenagers math so he knew how to make complex ideas land.
“I didn’t want to do a little HTML course,” he said. “I wanted to make the course that everyone would want to buy if they wanted to get into web development.”
He spent five months building it. His target: $3,000 to $4,000 a month. He would have been thrilled with that.
From First Step Failure To Launching For Free
Five pain-staking months later, Rob launched The Complete Web Developer Course at $199.
Day one – one sale – and the buyer requested a refund!
He didn’t panic though. Instead, he decided to make the course free to get some traction.
Most people would baulk at this move. A thousand signups, even at a low price of $10, is still $10,000.
But Rob understood something about marketplace platforms: a course with zero students and zero reviews doesn’t sell, no matter how good it is. Social proof comes first. Revenue comes after.
Within two weeks, 4,000 people had signed up.
Most barely watched anything. But the course page no longer showed zero students.
Reviews were harder. He called family. Emailed friends. Asked them to watch the first few videos and leave honest feedback. Twenty reviews later, the course looked real.
Then he did something most new course creators can’t do. He emailed his hosting customers.
“There was real synergy between my web hosting customers and this course,” he said. “A lot of them were small business owners. They didn’t all have coding skills, but they were all working with a website every day.”
That single email promotion brought in $15,000 in his first paid month. Around $10,000 – two-thirds of it – came directly from the hosting list.
His goal had been $3,000 to $4,000 a month. He’d tripled it before Udemy even noticed him.
But Udemy did notice. A course selling well with strong reviews triggered their marketing machine. Homepage features. Email promotions. Paid advertising. The platform knew how to scale something already working.
Month two: $30,000. Month three: $50,000. By month four, he was earning $150,000 monthly – a number that would stabilize as his new baseline.
Within a year, The Complete Web Developer Course alone had generated over $1 million.
Free Hosting = Recurring Revenue
Here’s where most course creators would have stopped. A million dollars from one course is enough to justify the work.
But Rob had bundled something into the course that no other Udemy instructor could offer: a free year of web hosting through Eco Web Hosting, valued at £119.
“I combined it with offering the web hosting, which I thought was a really nice extra that no one else would be able to provide,” he said. “Because most online instructors don’t have web hosting businesses.”
Students learning to build websites need somewhere to put them. Rob’s course didn’t just teach the skills—it gave them a place to practice for free.
What he didn’t anticipate was what happened next.
“It actually funneled one into the other,” he said. “The web hosting is now growing quite quickly because of the courses.”
Most creators on Udemy sell a course and move on to the next launch. The course revenue is a one-time payment. Once it’s done, you’re back on the treadmill, either launching something new or watching revenue decline.
But in a move of accidental genius, Rob had built something quite different: a system where course buyers became hosting trials, and a percentage of those trials converted to paying subscribers.
One-time revenue is a treadmill. You stop launching, you stop earning. Recurring revenue is a floor. It pays you while you build the next thing (or while you take a month off).
The two businesses now ran as one operation. Five employees. No separation between hosting and courses. Each business fed the other.
The $200,000 Month
By September 2014, three months after his first course went live, something unexpected happened.
Apple released Swift, a new programming language for iOS development. Udemy saw the demand spike and contacted Rob directly.
Would he be interested in building a Swift course? If he did, they would market it aggressively.
That invitation from the platform itself, reaching out to him rather than the other way around, represented something important.
Rob had built platform capital. His track record with the web developer course meant Udemy was now invested in his success.
He used Kickstarter to pre-validate the Swift course, aiming for a few thousand dollars in pre-sales. He raised $50,000.
When the course launched on Udemy in September 2014, it generated $200,000 in its first month, more than the web developer course had done initially.
Within a year of his first Udemy upload, Rob had four courses, 125,000 students, and nearly $3 million in gross revenue. After Udemy’s cut, roughly $1 million went to him.
Half A Million Students Want More
Success created a new constraint: students wanted more courses than Rob could build.
“The courses take a lot of time to put together,” he said. “Maybe a month or two for a typical course. And they don’t stay very static for very long. They need pretty regular updating.”
Every time Apple released a new iOS version, it often broke everything in the previous course. A whole new course was required. Rob couldn’t keep up alone.
So he stopped trying.
Other instructors started reaching out, asking to partner. Rob could provide the audience (over 500,000 students by this point) and his understanding of what made courses sell on Udemy. They would provide expertise in topics he’d never cover himself.
“We’re acting more or less like a traditional book publisher,” he said. “Finding someone who’s an expert, training them up, making sure the content is good, and helping with publicity.”
He called it Codestars. Today it has taught over 2.5 million students across 34 courses.
While Rob hasn’t disclosed Codestars-specific revenue, the publishing model meant he earned revenue from courses he never had to create, free from the capacity constraints that would have capped his solo income.
What You Can Actually Use
Rob Percival’s specific advantages – an existing hosting business, platform timing in 2014, and a teaching background.
These aren’t directly replicable, but the architecture is.
Solve distribution before content. Rob didn’t launch on his own website where nobody would find him. He used Udemy’s existing traffic to solve the discovery problem, then focused entirely on course quality. If you have no audience, consider launching on a marketplace first. Treat the lower margins as customer acquisition cost.
Manufacture social proof before revenue. Making the course free for two weeks looked like leaving money on the table. It was actually buying the reviews and student count that made later sales possible. Prove conversion with a smaller audience before scaling.
Launch to existing assets first. Two-thirds of Rob’s first $15,000 came from his hosting email list – customers who already trusted him. Before relying on platform algorithms, identify any existing audience or customer base you can promote to. Even a small list of high-trust contacts converts better than cold platform traffic.
Bundle something with recurring revenue. The courses were one-time payments. The hosting was monthly. By bundling them together, Rob converted a percentage of every course sale into ongoing subscription revenue. What adjacent service could you bundle with your product that creates ongoing relationship and ongoing revenue?
Partner when you hit capacity. Rob didn’t try to scale by working harder. When demand exceeded his ability to deliver, he became a publisher instead of a creator. Find domain experts who lack audience. Provide distribution and production methodology. Share revenue. You become the platform.
A decade after that afternoon in Ben’s office, Rob had built and sold two companies. Eco Web Hosting went to Enix. Blutick, an AI math tutoring platform he’d created, sold to AQA Education. He’s now building Learn Anything, an AI essay grader backed by the UK Department for Education.
His original goal had been $3,000 to $4,000 a month teaching web development.
Instead, he built a system where distribution solved discovery, recurring revenue created a floor, and partnerships scaled beyond his personal capacity.
The lesson isn’t “teach coding on Udemy.” The lesson is: solve your distribution problem first, build something that keeps earning after the initial sale, and know when to become the platform instead of the creator.