6 Months, 1500 Hours, 0 Paying Users: What I Learned After Abandoning Two Products

In June 2025, I shut down the Codofly development server.

6 months. 1000+ hours. 15,000 lines of code.

Result? 0 paying users.

Two weeks later, I abandoned Scholar AI.

Back then, I sat at my computer, staring at the carefully crafted code, feeling like a failure. I thought making good products was enough. I was wrong.

But these two failures taught me more than any success ever could.

If you're also building side projects, if you're struggling with why no one is using your product, this article might help you.

Product #1: Codofly - The Technical Trap

I'm a full-stack developer. I previously contributed to Sealos, which now has 15.6k stars.

So I thought I understood technology. I wanted to create an AI SaaS template to help developers launch products quickly. Sounds good, right?

In January 2025, I started writing code.

What I Invested

For the first two months, I spent significant time on architecture. Next.js 15, TypeScript, Prisma, PostgreSQL. I studied various best practices.

Authentication system? I used NextAuth.js 5, spending 2 weeks optimizing various edge cases.

AI integration? I supported OpenAI, Anthropic, Google, and xAI. Streaming responses, cost calculation, error handling - I considered everything.

Payment system? Stripe subscriptions, credit billing, Alipay, WeChat Pay - I had everything.

After 6 months, I had a "perfect" product. Clean code, elegant architecture, comprehensive features.

I'd give myself an 8/10.

Then I launched.

First Week

Visits: 47

Registered users: 3 (all my friends)

Paying users: 0

I thought, maybe it wasn't enough promotion. So I posted on Twitter.

No response.

Second Week

Visits: 31

New registrations: 1

Paying: 0

I started panicking. I posted on Reddit with the title "Brand new AI SaaS development template, integrated with multiple AI models."

15 reads, 2 replies, both asking "What's the difference from other templates?"

I couldn't answer. Because technically we were similar, or even mine was better.

But that wasn't the point.

My Three Fatal Mistakes

Mistake 1: I thought technology was a moat

I spent a full 2 weeks optimizing database query performance. From 200ms down to 50ms.

After launching, I discovered: nobody cared.

Users asked: "Can this help me make money?" "How long to launch?" "Do you have success stories?"

Not: "Are you using PostgreSQL or MySQL?" "How fast are the queries?"

I later studied ShipAny, an AI application template developed by idoubi.cv. His code might not be as "elegant" as mine, but he solved users' core problem: quick launch.

His slogan is about enabling developers to quickly build and deploy AI applications.

Mine? "Enterprise AI application development template, integrated with multiple AI models."

Which is more compelling?

Truth 1: Technology is only 30%, marketing is 70%

This was the most expensive lesson I learned in 6 months.

Mistake 2: Closed development, zero marketing

During the 6 months of development:

My Twitter updates: 0
My blog posts: 0
My community operations: 0

I thought "good wine needs no advertising." I thought good products would sell themselves.

Then I saw idoubi.cv's data:

Twitter followers: 20k+
Shares development progress daily
ShipAny continuously gains users
Actively participates in developer community

Me?

Twitter followers: 200
Newsletter: non-existent
Sharing frequency: never
Public data: never
Launch results: $0

The gap isn't in the product, it's in the brand.

Truth 2: Personal brand > product itself

If nobody knows who you are = nobody trusts your product.

Mistake 3: Who is the target audience? I didn't know either

My Landing Page said: "Enterprise AI SaaS template."

What does "enterprise-grade" mean?

Is it for large companies? But large companies have their own development teams, they won't buy templates.

Is it for startups? But they have limited budgets and will choose cheaper options.

Is it for individual developers? But the term "enterprise-grade" scares them away.

Result: nobody was my user.

Compare ShipAny's positioning: focused on rapid AI application development, lowering technical barriers

Clear. Practical. Target audience is clear.

MkSaaS: "One-stop AI application development platform," developed by indie_maker_fox

One-stop service? That promise is attractive.

My positioning? Vague, cold, technical.

Truth 3: If you serve everyone, you serve no one

Product #2: Scholar AI - The Passion Test

After Codofly failed, I told myself: it's okay, let's try again.

This time I chose a "bigger market" direction: academic research AI assistant.

Sounds promising, right? The academic community has so many people, there must be demand.

I spent 3 weeks doing market research, reading various research papers. Then started development.

One Month Later

I found myself completely losing interest.

Why?

Because I'm not the user.

I don't do academic research. I don't read papers. I barely even visit arxiv.

I just thought "this idea sounds good."

But you can't build products based on "sounds good."

Every time I wanted to add new features, I had to ask others: "Would academic researchers need this?"

I didn't know. Because I'm not them.

Then I thought about my other product: DevToolKit.

That's a developer tools curation website. Why did it succeed (now 1.4K monthly visits)?

Because I search for developer tools every day.

I'm the first user. I know where the pain points are, I know what content is valuable.

Every feature of DevToolKit exists because I needed it myself.

Scholar AI? I was building a product for a user group I didn't understand.

One month later, I abandoned it.

Truth 4: Scratch your own itch

The best products often come from your own needs.

The 5 Harsh Truths

Looking back, these two failures gave me 5 clear lessons.

Truth 1: Technology is only 30%, marketing and branding are 70%

My Codofly:

Code quality: 8/10
Marketing ability: 1/10
Result: failure

ShipAny:

Code quality: 7/10 (probably)
Marketing ability: 10/10
Result: continuous user growth and revenue

This comparison woke me up.

Good code is important, but it's not the decisive factor for success.

idoubi.cv's words made me understand: personal brand is more important than the product itself.

I thought this was motivational fluff back then. Now I know it's truth.

Truth 2: Community is the moat, not technology

Codofly's technical barrier: high (15,000 lines of carefully optimized code)

Codofly's community size: 0

Codofly's moat: non-existent

MkSaaS's technical barrier: medium (similar to other templates)

MkSaaS's community: active user base and developer community

MkSaaS's moat: strong (community helps with marketing, provides feedback, supports new users)

Technology can be copied. If you're ahead today, others can catch up tomorrow.

But community can't. An active, valuable community is the hardest asset to replicate.

I spent 6 months optimizing code, but ignored the most important thing: building community.

Truth 3: Rapid validation > perfect product

My way:

6 months of closed development
Pursued perfection
Only validated after launch
Found nobody wanted it
Time was wasted

The right way (what I learned now):

3-7 days to build MVP
Launch immediately
Collect data and feedback
Continue if growing, abandon if not
Time is the most precious resource

I've seen many successful indie makers do this.

Pieter Levels built 100+ products, only a few succeeded. But his validation cycle for each was short.

Me? Put all eggs in one basket, invested 6 months, and the basket was broken.

Truth 4: Building in Public is not an option, it's essential

I studied all successful SaaS boilerplates:

ShipAny (idoubi.cv): active community participation
MkSaaS (indie_maker_fox): continuous product updates and community building

They're all doing one thing: Building in Public

It's not about sharing only after success, but from day one:

Share revenue numbers (even if $0)
Share failures and setbacks
Share what you've learned
Share product decisions

Why is this so important?

Because transparency = trust = growth

When you share the real process openly, people will:

Trust you (you're not hiding)
Support you (witness your growth)
Recommend you (feel involved)
Become your community

I developed in private for 6 months, missing all these opportunities.

Truth 5: You're not selling a product, you're selling transformation

Look at successful product positioning:

ShipAny: focused on rapid AI application development and deployment → They're selling efficiency and technical convenience

MkSaaS: one-stop AI application development platform → They're selling complete solutions

My Codofly: "Enterprise AI application development template" → I'm selling... code?

People don't buy feature lists. They buy:

Dreams (what I want)
Fears (what I'm afraid of)
Transformation (from current state to ideal state)

ShipAny users aren't buying technical templates, they're buying the possibility of "quickly build AI applications, save development time."

Me? I was selling tech stack.

This is the gap.

A New Beginning

What did these two failures cost me?

7 months of time
1500+ hours
Countless late nights

But what did I gain?

Clear understanding
New strategy
These 5 harsh truths
Knowing what to do and what not to do

Now, I'm starting over.

But this time it's different:

✓ I'll build in public from day one

✓ I'll build my personal brand first, then push products

✓ I'll validate quickly, 3-7 day MVPs

✓ I'll build products I need

✓ I'll share real data, including failures

I set a goal for myself: 90 days, from $0 to $3K MRR.

Might not reach it, but that's not the point.

The point is the process. Is transparency. Is authenticity. Is growth.

If you're also building side projects, if you've also experienced failure, I want to tell you:

Failure isn't scary, not learning from it is.

I'll document every step of these 90 days. Real data, real struggles, real growth.

Welcome to follow my journey.

About Me

I'm zjy365, full-stack developer, Sealos contributor (15.6k⭐).

Currently operating:

DevToolKit (developer tools curation, 1.4K UV/month)
LinkedIn Formatter (LinkedIn formatting tool)

Starting today: 90-day Building in Public experiment

Find me:

Twitter/X: @zjy365
Website: zjy365.dev
GitHub: github.com/zjy365

Next week, I'll share my complete 90-day plan.

Let's witness this experiment together.