A WeChat Mini-Program Made $1.46 in 15 Days — and That’s the Point

Qian Xiao Qian · Half-Month Review

---

Let's start with the conclusion: In 15 days since launch, the mini-program + official account earned a total of 10.53 yuan.

Don't laugh. I know this amount can't even buy a cup of milk tea.

But what's interesting is—0 yuan on the first day, 0.30 yuan on the second, a sudden spike to 2.66 yuan on the eighth day, and then a decline. What happened in between?

I want to seriously discuss the story behind these 10.53 yuan. The complete journey of a programmer who doesn't understand product, can't design, and has never written a mini-program, using AI to go from zero to launch, and then to generating the first income.

If you are also hesitating about "whether to make your own product," this article might give you a little courage.

---

Data First

Income details from traffic monetization over 15 days since launch (unit: yuan):

Date	Mini-Program	Official Account	Total
8/10	0.00	0.00	0.00
8/11	0.28	0.02	0.30
8/12	0.07	0.06	0.13
8/13	0.07	0.03	0.10
8/14	0.22	0.01	0.23
8/15	0.02	0.02	0.04
8/16	0.00	0.05	0.05
8/17	0.18	0.03	0.21
8/18	0.73	0.07	0.80
8/19	2.53	0.13	2.66
8/20	1.88	0.06	1.94
8/21	2.24	0.04	2.28
8/22	1.00	0.06	1.06
8/23	0.80	0.01	0.81

Total: 10.53 yuan (Mini-program 9.94 + Official Account 0.59)

---

How Did These 10.53 Yuan Come About?

I created a product matrix called "Qian Xiao Qian"—a WeChat mini-program + official account, bringing the thousand-year-old Guanyin divination lots into the phone.

On the mini-program side, users shake their phone once a day, randomly draw a lot, and see the lot text and interpretation. On the official account side, a daily push of lot interpretations. The income comes from WeChat traffic monetization ads.

There are two noteworthy points in the income curve:

First, on 8/19 it suddenly spiked to 2.66 yuan. No promotion was done that day; it was purely organic traffic. This shows the product itself has some self-propagation ability—someone drew a good lot and shared it with friends.

Second, it then started to decline. By 8/23, it was only 0.81 yuan. No continuous influx of new users, and old users left after drawing their lot.

What does this indicate? The product has organic growth, but hasn't found a method for sustained user acquisition.

---

One Person + AI, 35 Days from Zero to Launch

Timeline:

• 7/20: Project initialization, first line of code
• 7/23: Mini-program frontend + backend basic functions completed
• 8/10: Backend deployment online, traffic monetization activated
• 8/24: Today, income 10.53 yuan

From idea to launch, 35 days. From launch to generating income, 1 day.

But I want to emphasize—this is not a story of "a tech god grinding out a product in 35 days."

This is a story of an ordinary programmer + AI collaboration. And there were quite a few stumbles along the way.

---

The Real Process of AI Collaboration

Many "AI-made products" articles online read like a credit list for AI: AI designed the database, AI wrote the frontend, AI did the UI... as if a person only needs to say a word and the product appears.

The real situation is not like that.

Database Design: AI gave a plan, I rejected it three times

When building the achievement system, AI initially stuffed all information into one table—unlock status, claim records, shipping addresses all together. I said "No, claim records should be a separate table; there might be other scenarios in the future that also need to claim gifts." AI changed it, but added a claimed_at field. I said "Not needed, created_at already exists." AI changed it again.

After three rounds of back-and-forth, the final plan settled: one table only records "whether unlocked," another general gift claim table uses source_type + source_id to support multiple sources.

AI provided the starting point, but the decision must be made by a human.

Frontend Development: Request, review, request changes, iterate again

The mini-program's interface wasn't generated by AI in one go. The glassmorphism style, bamboo stick animation, achievement card layout... each version involved me first looking at the effect, then saying "this spacing is too large," "the color is wrong," "change the layout," and AI would revise it.

Sometimes AI over-designs—adding a bunch of features I don't need. I had to learn to say "don't want this."

Content Operations: 100 lot interpretations, AI wrote the drafts, I edited every single one

For the interpretation articles of 100 Guanyin lots, AI could indeed generate them in batches. But the generated content was monotonous, reading like a template. I spent a lot of time adjusting the tone, adding a personal perspective, giving each lot interpretation its own flavor.

AI handles quantity, humans handle quality.

UI Design: Concept is human, code is AI

The glassmorphism style, the visual representation of bamboo sticks, the interaction of achievement cards... these design decisions were all made by a human. AI's role was to translate design ideas into code.

---

What Do These 10.53 Yuan Signify?

It cannot signify that the product is successful.

But it can signify a few things:

First, one person can indeed make a complete product. No team needed, no financing needed, no need to wait until "ready" to start. 35 days, from 0 to a usable product that people use and that generates income.

Second, AI greatly expands the capability boundaries of an individual developer. Work that previously required four people—frontend, backend, design, operations—can now be covered by one person + AI. But the premise is—you must know what you want and be able to judge whether the solution AI provides is good.

Third, traffic monetization income is real, but also brutal. 10.53 yuan in 15 days, an average of 0.70 yuan per day. At this rate, about 21 yuan a month, 252 yuan a year. Server costs are about 500 yuan a year, not even halfway to breaking even.

But it is a breakthrough from 0 to 1. Going from "no one uses it" to "someone uses it, there is income"—this qualitative change is far more important than the number itself.

---

Next Plans

• Optimize ad placement layout to improve eCPM
• Explore linked customer acquisition between the official account + mini-program
• Continuously iterate on product experience

(The achievement system and lot collection gameplay are already online; retention effects are being verified.)

The goal is simple: first, get the average daily income to exceed 1 yuan. At the current growth rate, it will probably take... a long time.

But every step is real.

---

Final Words

I am not advocating that "using AI to make products will make you money."

The income of 10.53 yuan shows that making a product is easy, making a product people use is hard, making a product that can sustainably make money is even harder.

But I also want to tell you: If you have an idea, don't wait any longer.

AI has lowered the barrier for individual development to a historic low. You don't need to be a full-stack engineer, don't need to be a designer, don't need to be a product manager. You only need to—start doing it.

For the rest, AI will help you. But the direction, you must set yourself.

---

Qian Xiao Qian, a Guanyin divination lot mini-program made with AI. Scan the code to experience 👇

Follow the official account for daily pushes of lot interpretations 👇