Ben Stokes's Mailoji: Turning Small Experiments into a Business

Ben Stokes is a developer who has built small internet products and shared the process on ‘Tiny Projects’. His best-known work, Mailoji, is a service that sells email addresses containing emoji, making personality and fun—rather than work efficiency—the product. What matters in his case is not one unusual idea, but the process of launching, operating, or selling multiple products while finding a business that suited him.

Turning Ideas Piled Up in a Notes App into Real Products

Tiny Projects began in May 2020. At the time, he was a full-stack developer who had experience with businesses including ice cream and cookie dough, and project ideas he had never executed kept piling up in his phone’s notes app. Prompted by his twenty-fifth birthday, he decided to build the ideas even in small forms, and used a blog as the record of that experiment.

The early product, One Item Store, was a simple online store-building tool that let users sell a single product repeatedly. He built it in about two weeks and took a 1% fee on sales, but the revenue he generated by running it himself was less than £2. However, after someone who read the build story proposed an acquisition, it was sold for $5,300 in late 2020. Even though the product’s operating revenue was negligible, there was a buyer who needed already-built software.

The next product, Earlyname, was a service that helped people secure short, rare usernames on new social services first. By adding a $10-per-month username reservation service, he grew it to $350 in monthly revenue over about six months, but the work of researching new services and sending a newsletter every month became burdensome. The workload was about five hours a month, but he did not want to keep running it, and he sold it for $10,500 in 2021. This experience revealed that ‘a product that makes money’ and ‘a product you want to keep running’ can be different.

Mailoji, Which Started with 300 Emoji Domains

Mailoji's starting point was curiosity about whether emoji domains could also be used as email addresses. He first experimented with 📪.ws, shaped like a mailbox, and while looking for other emoji, arrived at Kazakhstan's .kz domain. At the time, single-character emoji domains cost about $8 each, and he first spent $1,200 on 150 of them. He later increased his holdings to 300, and the domain purchase cost reached about $2,400.

Mailoji was not a new mailbox to replace Gmail, but a service that forwarded mail received at an emoji address to an existing email. Users connected an address such as ben@🚀.kz to their Gmail or other email, and replies were sent from the email address they originally used. What was sold was therefore closer to a distinctive address attached to the front of an existing mailbox than to a complete email system. The official guidance also distinguishes between sending directly from an emoji address and forwarding received mail; Mailoji forwards received email but does not let users send directly from its emoji address.

The first version was built in a few weeks by combining HTML, CSS, JavaScript, and the Stripe payment API. The early problem of emails not being delivered properly was solved by experimenting with a relay structure using AWS. The product scope was focused not on an entire new mailbox but on address registration and forwarding.

Three Public Launches That Turned Attention into Sales

The first customers came from TikTok. Unable to run paid ads because he could not submit the VAT registration number required during the ad account application process, he posted the video he had prepared as a regular post. The video was viewed more than 200,000 times over two days, and about 60 email addresses were sold.

Then, on Product Hunt, it rose to No. 5 on launch day, and more than 150 addresses were sold. However, the annual recurring revenue (ARR) disclosed in an early retrospective written on March 11, 2021, was about $1,440, and the domain purchase costs had not yet been recouped. There was an initial response, but it was not yet at a stage where profitability had been validated.

The big change happened after he posted the build log to Hacker News. In a May 2021 post, he said that Mailoji recorded about $9,000 in sales over one weekend. An Indie Bites interview published on June 14 of the same year put “$10,000 over a weekend” in its title, but in its introduction explained that Mailoji had just passed $10,000 in sales. Therefore, it is more accurate to distinguish between the “approximately $9,000 in weekend sales” in his own written record and the “passing $10,000 in initial sales” in the interview.

This $10,000 is not $10,000 in monthly recurring revenue (MRR). Because Mailoji was a product that collected annual subscription fees upfront, one should not equate initial new subscription sales with cash that recurs every month. Conversely, explaining it only as sales of one-time products such as lifetime licenses is also inaccurate. In a January 2022 post, he stated that the first subscription renewal was due on February 1 of that year, meaning there was a gap of about one year between the initial sale and the actual rebilling.

There is also a record that disclosed recurring revenue metrics separately. On September 22, 2021, Mailoji’s ARR was about $13,000, and the number of selectable emoji domains had grown to about 700. He stated that the annual subscription revenue generated by the single most popular one, 🚀.kz, was about $1,700. Not all domains sold equally, and certain emojis generated revenue in a concentrated way.

In the process of expanding features, he also received help from an outside developer. To implement the feature that allowed emojis not only in the domain of the email address but also in the part before the @, he hired a developer on Upwork and built a custom email server. It is true that it was an individual-led product, but presenting it as a case in which he implemented all the technology alone from start to finish does not match the actual process.

The explosive sales immediately after launch did not continue. In November 2021, it normally sold about one annual subscription per day, and the regular price was $10 per year. When it was discounted to $5 per year for Black Friday, sales increased to about 20 per day; afterward, operations were closer to steady small-volume sales and occasional promotions than to large-scale virality.

The Operational Weight Revealed After the Sale

The strategy of securing many domains came with ongoing costs. He thought that rare emoji domains increased the product’s value and made it difficult for competitors to enter, but he was also wary of the fact that renewal fees were incurred every year. In fact, because he was reinvesting the proceeds from selling other projects and Mailoji subscription sales into buying domains, the disclosed sales amount cannot be viewed directly as his personal net profit.

In January 2022, an unexpected external risk became reality. During Kazakhstan’s internet shutdown, he recorded that 300 of the .kz emoji domains he owned stopped working, and he could not access the website of the local domain registrar either. The email service was later restored, but a business selling fun addresses was also not free from disruptions in domain and communications infrastructure.

The first subscription renewal also brought payment failures. He had added emails in advance notifying customers of upcoming renewal dates, card expiration, and payment failures, but in February 2022 he experienced a high card decline rate on repeat payments from customers initially acquired through TikTok. Even after asking Stripe to investigate, he could not determine the cause, and he responded by adding alerts that notified him of the reason for failure. Selling a subscription product and retaining that subscription into the following year were separate operational challenges.

In March 2022, about 800 subscriptions generated by the previous year's Hacker News traffic reached a major renewal point. He reported that although there were some cancellations and payment failures, many customers remained, and that the churn rate at the time was about 20%. This is a figure he presented when explaining the outcome of that renewal; it does not mean that monthly churn was 20%. It is an important follow-up record in that customers continued to use the service even after the initial novelty wore off.

In a retrospective on May 18, 2022, the cumulative revenue he disclosed for Mailoji was $18,500. It was not a project that ended with a weekend sales spike; it continued as a service whose revenue accumulated through the following year's renewals. However, this too is cumulative revenue disclosed by the founder, and it does not mean net profit or monthly revenue.

How Small Experiments Create the Next Business

Not every experiment at Tiny Projects made money. According to a 2022 retrospective, an 8-bit battle royale game had 183 downloads and no revenue, and Snormal, a social service for posting ordinary daily life, was likewise left effectively abandoned with no revenue. If you isolate only the successful Mailoji, it may look as though a clever idea immediately leads to revenue, but the actual record also includes products that drew no response or were discontinued.

He himself acknowledged the limits of making multiple products. In 2022, he was running three active projects largely automatically, but he thought that once there were five or more, time management and switching between tasks could become difficult. He also wrote that as his readership grew, the temptation grew to build products that would receive many recommendations online rather than products he actually wanted to operate. Building in public was a means of bringing in customers, but it also risked turning product selection into a popularity contest.

Meanwhile, after Mailoji, he built Paper Website, which turns handwriting into a website. The method was to photograph writing on paper, extract the text through character recognition, correct recognition errors with GPT-3, and then publish it to the web. He also used this product for his own daily blog, continuing an approach of combining existing technologies to realize the small user experience he wanted.

The problems he encountered while building Paper Website led again to the next product. From the experience of spending time making his own prompt because he could not find a good prompt to correct character-recognition errors, PromptBase, a marketplace for buying and selling AI prompts, emerged and was launched in June 2022. The 2026 Tiny Projects homepage states that this eighth project had grown to become his main business. He had not decided on a big business from the start; rather, he discovered the problem for the next business while building the previous product.

Ben Stokes's case is closer to a process of deciding what to learn from and what to keep from each product than to launching multiple products itself. He sold products that did not fit his operations, kept products that could be maintained with relatively little management, and focused his time on products where a larger opportunity appeared. Mailoji's initial $10,000 was a notable achievement in that process, but the longer-lasting asset was the ability to experience actual sales, renewals, and incident response and connect that to the next product.