Codotx First Year in Business: A Candid Review
Here is the bottom line: the money ran out, the products weren’t selling, and next year we need to take on client projects to stay afloat. While none of the products succeeded, we accumulated some valuable failures — learning which paths to avoid, and discovering entirely new landscapes and opportunities after diving deep into AI-powered development.
This Year’s Harvest
Plugin sales generated NT$280,000, averaging NT$23,000 per month, all from a single plugin’s revenue. Naturally, the thought was: if there were a second or third plugin, earning NT$100,000 per month wouldn’t be a dream. So at the start of the year, we began developing a second product.
The first product did well with customer order notifications, so the second one continued the theme but focused on admin notifications, integrating AI to let administrators query orders using natural language. The business model was to release a free version first, then charge for different AI query features.
However, during testing, the AI query functionality proved unstable. WooCommerce version differences caused incorrect data — every time you asked for this week’s revenue, a different number appeared. Making it compatible across versions was an enormous engineering effort, and with AI output being hard to control at the time and no clear business model, we decided to hit the brakes.
A Full-AI Social Platform Experiment
Next, we tried building a social platform entirely with AI, choosing React and Supabase — technologies we were completely unfamiliar with — developing everything from scratch using only AI. I only adjusted the CSS myself. After launch, not only did no one pay, there weren’t even any active users.
Looking back at the failure: we vastly underestimated the influence of existing social platforms. We built a weak-tie community that was destined to have no users from the start. Even more dangerous, AI Coding kept us in a fantasy bubble — when AI constantly praises your ideas, it becomes even harder to see reality.
Returning to the WordPress Ecosystem
The fourth product returned to our roots, focusing on LINE customer service integration. This was a collection of features many clients had requested, and also problems we encountered ourselves while running a LINE Official Account. This time, we wrote articles first to validate market demand before investing in development, only starting after collecting enough survey responses.
The first version took one month to complete, and four customers placed orders on launch day. But those four orders turned out to be the sales ceiling — after one month on the market, not only did existing customers cancel subscriptions, but no new customers came in either. When asked about cancellation reasons, responses were either silence or that managing through the LINE dashboard was more convenient.
For businesses, the switching cost from LINE’s dashboard is much higher than expected. Eventually, integrating AI chatbot functionality showed a turning point — new customers started ordering for the AI customer service feature, and existing customers began trying it again.
One Year of Insights
Entrepreneurship is no longer just textbook knowledge — it’s a path walked firsthand. After understanding our own shortcomings, seeing others who survive on the entrepreneurial path fills us with genuine respect. Next year’s direction is clear: stabilize income with client projects while continuing to refine products, waiting for the next breakthrough. Thank you to all our customers and partners who have supported us.