Ordinary Man Trying

I Built This Blog in 4 Days Using AI — And Now I’m Running a Live Experiment in Public (Follow Along)

I Built This Blog in 4 Days Using AI — And Now I’m Running a Live Experiment in Public (Follow Along)

This post is pinned and updated continuously. Think of it as a live reality show — except the budget is ¥700 (~$100) and everything is real. Come back anytime for the latest episode.

A Declaration Before We Begin

My child is in high school right now. Graduation is roughly five years away.

I have five years to figure this out — or fail trying — before they step into the world and have to figure things out themselves.

So I want to leave them something more useful than advice. I want to leave them a record. A real one. With the numbers, the mistakes, the dead ends, and the occasional small win.

I also built a 5-year voting experiment: which Chinese university majors will be most valuable in 4 years? China has 757 official university majors — I translated every one of them into English and let readers vote. If you have kids at that stage, or just an opinion about where the job market is heading, your vote counts. I’ll compare the results to reality in 2028.

This blog will keep updating for at least five years. Even with zero readers. Even with zero revenue. Even if Google never ranks a single page.

If you are reading this, you are watching a parent try to show a teenager what building something from nothing actually looks like — not the highlight reel, the whole messy process.

That is the only mission here.

Why This Exists

Four days ago, I did not have a blog.

I did not know how to register a domain. I did not know what web hosting was. I had never written a single word in English for a public audience.

I am an ordinary Chinese person. My English is functional, not fluent. My tech skills are basic. I have no followers, no network, no marketing budget.

I did all of this with AI guiding almost every step. I asked questions like a complete beginner — because I am one.

What I’m Up Against

Let me be honest about the odds.

AI is replacing jobs. The reason I’m building something independent is that I don’t know what my industry looks like in five years. Neither do you. Neither does anyone.

English is not my first language. I write every article with AI helping close the gap. Some readers will notice. I’m doing it anyway.

I’ve failed at this before. Two previous websites. Six years. Roughly zero dollars. My mom still asks me why I keep trying.

Google might never rank a single page. Most blogs never get meaningful search traffic. The majority of websites quietly die. I know this.

I’m restarting in my 40s. Not a 22-year-old with 50 years of compounding ahead. A middle-aged parent with a real job, limited hours, and a child watching.

These are the real conditions. I’m not pretending otherwise.

I’m doing it anyway — because the alternative is to let five years pass without trying.

📺 What You Are Watching

This is a slow-burn public experiment with one simple rule: nothing is hidden.

Every traffic number, every failure, every sale (or zero sales) gets posted here. You are watching a real website try to survive — Google, Amazon, and the open internet as judges.

If you want to help, the best things you can do are:

  • 👀 Come back — more readers = more reason to keep going
  • 📤 Share this post — one share could change the whole story
  • 🛒 Use any tool below and buy a book on Amazon — even $0.99 counts as a qualified sale toward our 180-day target

No pressure. Truly. But if you are rooting for this experiment, now you know how to help.

📅 Live Timeline — Every Episode Logged Here

Date Episode Key Metric
Jun 16, 2026 🟢 Day 1 — Site goes live. Domain registered, hosting connected, WordPress installed. Pages: 1 | Traffic: 0
Jun 17, 2026 ✍️ Day 2 — First articles published. First AI tools built and embedded. Pages: ~5 | Traffic: 0
Jun 18, 2026 🔍 Day 3 — Applied to Google Search Console. Sitemap submitted. GSC: pending
Jun 19, 2026 📊 Day 4 — Google Search Console approved. First impression recorded. GSC impressions: 1 | Clicks: 0 | Avg. position: 9
Jun 20, 2026 🛒 Day 5 — Amazon Associates approved. Book links added to 10 tools. PageSpeed 98/100. SEO 100/100. This post published and pinned. Affiliate links: 10 | Sales: 0 | Menu item added
Jun 21, 2026 📊 Day 6 — Google Analytics added to all 29 static tool pages. Open Graph tags added so tools show rich preview cards when shared on social media. World Cup AI experiment post entering its “results window” — predictions reveal June 24. Tools with GA4: 29 | Tools with OG: 29
Jun 21, 2026 🚨 Day 6 (continued) — Installed Sassy Social Share plugin. Homepage layout collapsed entirely. Called in AI to diagnose and fix. Root cause: Hostinger theme doesn’t load wp-block-library.css. Fix: all layout CSS rewritten as inline styles. AI created backup files, a Python update script, and a full maintenance guide. Homepage now AI-managed. WordPress theme updates paused indefinitely. Full story → Articles: 33 | Free Tools: 20 | Floating share bar: hidden | Comment section: redesigned | Tools nav added
Jun 21, 2026 🎓 Day 6 (continued) — Launched the Gaokao Major Selection Tool: all 757 official Chinese university majors translated into English, each with its own detail page (~1,000 words of AI career analysis, star ratings, FAQ schema). Built a PHP backend so votes persist server-side with IP rate limiting — one real shared leaderboard instead of per-browser localStorage. Added 13 ranking & category pages (Best Majors 2025, AI Picks, AI-Resistant, Engineering, Medicine, etc.). Total pages live: 771. Submitted full sitemap to Google + Bing/IndexNow. Full story → Pages: 771 | Votes: 120 | Backend: PHP
Jun 22, 2026 📂 Day 7 — GitHub strategy unlocked. 3 repos live on daligao.github.io. 6 new backlinks from GitHub domains. Repos: 3 | Backlinks: +6
Jun 23, 2026 🐞 Day 8 — Fixed 3 bugs in AQ Kids (one root cause). Added Ko-fi tip button — first payment option ever on any tool. Ko-fi: live | Bugs fixed: 3
Jun 28, 2026 📊 Days 9–14 — 307 active users this week. Loneliness Test shipped. Claude token limit hit mid-build. Rebuilt with 30+ Gemini iterations. Users: 307 | New tool: 1 | Revenue: $0
Jul 1, 2026 ✍️ Day 15 — Claude back. 3 articles published. Dev.to crosspost live. YouTube comment eaten by spam filter. Articles: +3 | Dev.to: 1 | YouTube spam: 1
Jul 12, 2026 🗺️ Days 21–27 — Wander China launched (17 cities). Changsha & Shanghai city pages deepened with real local detail. Amazon affiliate added (Lonely Planet China). Users: 57 | New tool: Wander China | Revenue: $0
Jul 15–19, 2026 📈 Days 30–33 — Traffic doubled to 143 users. Avg engagement hit 1:32. Wander China Planner became the #1 page (132 views). Microsoft Copilot cited the site 9 times in 7 days. Users: 143 | Copilot citations: 9 | Amazon clicks: 15 | Revenue: $0
Jul 17–19, 2026 💑 Days 32–33 — Launched Relationship Lab: an interactive attachment style analyzer that maps your relationship patterns across 4 styles. Built with full mobile optimization. No login, no data stored. New tool: Relationship Lab | Revenue: $0
Jul 19–25, 2026 📊 Days 34–40 — AI Investment OS becomes #1 most-visited page (124 views). Published childhood home AI restoration article. Submitted to Google and Human Parts (Medium). Google avg position improved to 1.8. Bing Copilot: 11 citations. Active users: 172 | Copilot citations: 11 | Revenue: $0
Jul 26–Aug 1, 2026 🛠️ Days 41–47 — Built two new tools: Gaokao English Study App and AI Chinese Story World (first step into video content). Google position slipped from 1.8 to 4.4. 150 active users. Amazon and Ko-fi: $0. Active users: 150 | New tools: 2 | Revenue: $0
Aug 2–8, 2026 📊 Days 48–54 — Built Standup Timer and Website Security Checker. Bing delivered 8 clicks from 200 impressions — more than Google this week. 188 active users (+25%). External backlinks hit 360. Active users: 188 | Bing clicks: 8 | Revenue: $0
Aug 9–16, 2026 📊 Days 54–62 — Launched chinesefortunetools.com in 48 hours. New site topped GA4 with 82 views on its first week. 225 active users (+20%). Bing Copilot cited the site 65 times in 7 days. Bought new 4-year Hostinger plan at $114.82, made $45.93 back in affiliate cashback. Active users: 225 | Copilot citations: 65 | Sites: 2 | Revenue: $0
Next update coming… Follow along to see what happens next. —

📊 Current Site Stats (updated: Sept 20, 2026)

Metric Now Goal
Days online 97 1,825 (5 years)
GA active users (this week) 198 Growing
Avg engagement time 60s 2+ min
Google avg position 7 (0 click from 2 impressions) First organic click
Bing AI citations (7 days) 624 citations, avg 3 pages/day Growing
PageSpeed score (mobile) 98 / 100 ✅
Amazon sales (180-day window) 0 / 3 needed (0 clicks this week) Deadline: Dec 17, 2026
Total revenue earned $0.00 Anything > $0
Active sites 4 (ordinarymantrying.com + chinesefortunetools.com + chinesenamecraft.com + chinarules101.com) Growing
Years committed to this 5 Until my child graduates

🛠️ The 10 Tools With Amazon Book Links

Each simulator below ends with a real book recommendation. If you play through and click to Amazon, any purchase helps keep this experiment alive.

📅 Day 7 — June 22, 2026

AI Taught Me to Use GitHub. I Almost Stopped at a Fake Paywall.

Reddit account still frozen (4 months now). Posts on X get 2 views. Asked AI: how do I get people to find my site?

AI said: GitHub.

Spent the day building repositories for 3 tools — Investment Punch Card, China Civil Service Challenge, 2026 Gaokao English Exam. Each README links back to this blog. Then AI showed me how to turn each repo into a live website with GitHub Pages.

Almost quit when I saw “Start free for 30 days.” Thought it was another subscription. AI explained: that’s only for private enterprise pages. Public GitHub Pages is free forever.

Clicked Save. Three tools now live on daligao.github.io. Six new backlinks from GitHub domains. Zero dollars spent.

New article published (draft): My Reddit Account Was Frozen. My Tweets Got 2 Views. Then AI Told Me to Try GitHub.

Repos built today:

📅 Day 8 — June 23, 2026

First Time I Asked to Be Paid. Also Fixed Three Bugs That Were Actually One Bug.

Spent the day on AQ Kids — the resilience quiz builder for parents and children. Launched it last week but never really “finished” it. Today I did.

The bugs: share link was over 2,500 characters long. QR code was a blank white box. Child’s “Continue” button looped back to the start of each character introduction. All three bugs had the same root cause — the URL encoding was bloated. Fixed the encoding (90% shorter URLs), and all three bugs disappeared.

Also spent an hour rebranding AQ Kids to “AQ Sandbox / Historical Decision Simulators.” Then changed it all back. The original name is clearer. Lesson: rename things only if the new name is obviously better, not just different.

The real milestone today: added a Ko-fi tip button to the quiz results screen. First payment button I’ve ever put on anything I’ve built. If someone taps it after their child finishes the quiz, that’s the first revenue from AQ Kids.

Also rewrote Step 3 (where parents write a message for their child) from “Write Your Message” to “Write a Letter to Your Child” — a small change that makes the whole flow feel different.

New post published: I Spent a Day Fixing Bugs on a Kids Quiz — Then Added a Tip Button for the First Time

AQ Kids live: ordinarymantrying.com/tools/aq-kids/

📊 Weekly Update: The “Cold Start” Reality (June 22-28)

Building in public means sharing the absolute truth. Here are the raw numbers for a site starting from zero:

  • Traffic: 307 Active Users. Average engagement time is just 28 seconds—I need to work on capturing attention faster.

  • SEO: 107 Impressions and exactly 3 organic clicks. The Google Sandbox is real, and SEO is a waiting game.

  • Revenue: $0.00. Generated exactly 2 clicks on Amazon affiliate links.

Embracing the “$0 phase” is tough, but it’s a necessary rite of passage.

🚀 New Tool: The Loneliness Test

You can’t control the algorithm, but you can control what you ship.

New Release: The Loneliness Test — Social Icebreaker

It’s a 13-level interactive game testing how much solitude you can handle. But its real secret? It automatically generates a fun, low-pressure invitation text so you can finally ask that friend out without the awkwardness.

Let’s see if shifting from “rational calculators” to “emotional icebreakers” drives more organic sharing.

📅 Day 15 — July 1, 2026

Claude Came Back. I Published 3 Articles and Got My YouTube Comment Eaten by the Spam Filter.

Short version of the past week: Claude’s weekly token limit ran out on Friday night, mid-project. I lost access for three days. I tried Aider + DeepSeek (got trapped in an infinite SEARCH/REPLACE loop for 40 minutes, zero code changed). Then I iterated 30+ times in Gemini’s chat to finish the Loneliness Test. Then I wrote an article about the whole experience. Then Claude came back, and we caught up on everything.

What shipped this week:

YouTube comment experiment — failed cleanly. Left a genuine comment under a loneliness TED talk with a link to the Loneliness Test. YouTube’s spam filter ate it silently. Lesson: never put an external link in a first comment on a large channel. Engage first, link only in profile bio.

Traffic this week: 1–3 users per day. One came from Warsaw. Bounce rate: 100%. The site is invisible to search engines right now — normal for a 15-day-old domain. Google needs time. We wait.

Revenue: $0.00. Days running: 15. Giving up: not on the agenda.

📅 Days 21–27 — July 7–12, 2026

Wander China Went Live. Real People Used It. The Numbers Don’t Lie.

This week’s milestone: Wander China went live. 17 cities. AI-generated 1-day routes. A Changsha deep-dive with local food, scam warnings, and foreigner-friendly tips that took three sessions to build properly.

Then something unexpected happened. The tool showed up in this week’s top traffic pages — before any promotion, before any SEO ranking, before I even announced it. Changsha Travel Guide pulled 12 views. The Wander China main page pulled 10. The Trip Planner got 11. Three pages from one tool, all in the top 10, in week one.

I don’t know where the traffic is coming from. Direct? A share? A bookmark? Google Analytics doesn’t say. But someone is using it.

This week’s numbers:

  • Google Analytics: 57 active users, 76 new users, avg engagement 39 seconds, 701 events
  • Google Search Console: 0 clicks, 2 impressions, avg position 1.0 — the site is barely visible in organic search yet
  • Amazon affiliate: 0 clicks this week
  • Top page: Homepage (69 views) — people are landing and exploring
  • Surprise: AI Decision Toolkit (67 views, only 6 unique users) — same 6 people kept playing it

The honest read: 57 users sounds small. For a 4-week-old blog with zero promotion budget and zero social following, it feels real. These aren’t bots. They’re staying 39 seconds on average. They’re clicking through tools. The Wander China tool and Changsha guide both cracked the top 10 without a single share or backlink.

Google Search Console shows almost nothing: 2 impressions total, avg position 1. That position-1 number is misleading — it means when my page appeared, it ranked first, but almost nobody searched for it yet. Organic search is a long game. I know this. Still playing it.

What shipped this week:

  • Wander China — AI China Trip Planner — 17 cities, 1-day AI routes, Changsha as the flagship deep-dive (local food, scams, budget tips, foreigner reality check). Built in one week from scratch.

Revenue: $0.00. Days running: 27. Still here.

Last updated: July 12, 2026 
This post is pinned. It will keep updating for 5 years.
— An ordinary Chinese person, trying.

📅 Day 29 — July 15, 2026

One Month In. I Read Every Number. Then I Made a Hard Call.

Today I ran a full one-month diagnosis across Google, Bing, and Yandex — all three search engines at once, with real data exports.

The headline: Google sent me 1 organic visitor in 28 days. Bing sent 33. Total clicks from Google Search Console: 3. I published the full breakdown as a separate article: Month 1 Website Self-Diagnosis.

The hard part: I finally admitted that the 757 English major pages I translated and published in bulk were almost certainly the reason Google pulled back. New site + 757 templated pages published at once = content farm signal. The impressions peaked on June 23rd then collapsed.

So today I made the call: removed both majors sitemaps from robots.txt and added Disallow rules for /tools/majors/ and /tools/majors-cn/. Those 757 pages will slowly drop out of Google’s index. It stings a little — that was weeks of work — but carrying them is worse than letting them go.

Also fixed 5 broken 301 redirects that Bing flagged. Bing was the only search engine that actually told me what was wrong. Google never said a word.

Gemini tells me the sandbox period is around 6 months. So I have until December to build something worth ranking. That’s the actual timeline now.

📅 Days 30–33 — July 15–19, 2026

Traffic Doubled. Copilot Cited Me 9 Times. Amazon: Still Zero.

Something shifted this week. The numbers are no longer embarrassing.

143 active users in 7 days. That is 2.5× last week’s 57. More importantly: average engagement time jumped from 39 seconds to 1 minute 32 seconds. These are not accidental visitors. They are staying, reading, clicking through tools.

The top page this week was the Wander China Trip Planner with 132 views — almost 20% of all traffic. Someone is finding it and sharing it. I still don’t know exactly where they’re coming from. Warsaw was the top city with 31 users. Then Boardman (12), Shinagawa City (8), Guangzhou (7), Paris (5).

The most surprising metric: this Building in Public post had 29 views and 4 minutes 2 seconds average engagement — the longest of any page on the site. People who find it actually read the whole thing.

The Copilot moment. Bing Webmaster Tools now shows a “Copilot Citations” panel. This week: 9 citations from Microsoft Copilot in 7 days, with a peak of 3 citations on July 15. That means when someone asked Copilot a question, it pulled content from this blog as a source. The site is 33 days old. AI search is indexing faster than human search.

Google Search Console: 0 clicks from 16 impressions. Avg position: 4.1. Position 4.1 is not bad — that’s first page. But CTR is 0%, which means the title or description isn’t compelling enough to click. Or the search intent doesn’t match what the page offers. I need to test different meta descriptions.

Amazon Associates — the hard number: 15 total clicks since launch. /bin/zsh.00 earned. 0 purchases. The 180-day countdown to earn 3 qualifying sales started December 17 — deadline is now roughly 5 months away. I need 3 purchases from 15 clicks. The conversion rate required: 20%. Normal Amazon affiliate conversion is 1–4%. This is going to be tight.

This week’s numbers:

  • Google Analytics: 143 active users, 145 new users, avg engagement 1 min 32 sec, 1,574 events
  • Google Search Console: 0 clicks, 16 impressions, avg position 4.1
  • Microsoft Copilot citations: 9 in 7 days (peak: 3 on July 15)
  • Amazon affiliate: 15 total clicks, /bin/zsh.00 earned, 0/3 qualifying sales
  • Top traffic source: direct (125 users), then GitHub (4), Bing organic (2)

Revenue: /bin/zsh.00. Days running: 33. Copilot citations: 9. Still here.

Built a Tool. It Became the #1 Page. Google Position: 1.8. Amazon: Still Zero.

This week’s main project: AI Investment Operating System — a free tool that walks through 5 investment decision steps with AI. I built it, published it, and by end of week it had become the single most-visited page on the site: 124 views, 35 active users. Nothing else came close. The homepage got 100 views. The tool beat the homepage.

That tells me something: tools get used, articles get read once. People come back to a tool.

Google Analytics week of July 19-25 2026 — 172 active users

GA4 this week (July 19–25): 172 active users (up from 143 last week, +20%), 174 new users, average engagement 1 min 43 sec (up from 1:32), 1,700 total events. The engagement time going up matters — people are staying longer, which means the content is getting better or the right people are finding it.

Google Search Console: 0 clicks from 15 impressions. Average position: 1.8. That position number is the best I’ve seen — up from 4.1 last week. Position 1.8 means the site is showing up near the top of the first page for some queries. But 0 clicks from 15 impressions at position 1.8 means either the queries are very low-volume, or the title/description isn’t compelling people to click through. Next thing to test: rewriting Yoast meta descriptions on the top pages.

Bing AI Performance: 11 citations from Microsoft Copilot in 7 days (up from 9 last week). Average cited pages: 1. The number is small but consistent — every week the AI is pulling from this site as a source. Week 6 it was 9, Week 7 it’s 11. Slow growth, but growth.

Amazon Associates: 14 clicks in the last 30 days. $0.00 earned. 0 qualifying purchases. The 3-purchase deadline is coming. I’m not going to pretend the numbers look good here — they don’t. 14 clicks across 30 days from a site with 47 tools and 44 articles is not enough traffic volume to convert naturally. This is the honest version of building in public.

Other things that happened this week: Published a post about using AI to restore photos of my abandoned childhood home (55 years old, empty for 20+ years). My mother cried when she saw the restored images. The post went live and I submitted it to Google for indexing immediately — first time I’ve felt genuinely confident about a piece of content on this site. Also submitted a version to Human Parts on Medium.

This week’s numbers:

  • Google Analytics: 172 active users, 174 new users, avg engagement 1 min 43 sec, 1,700 events
  • Google Search Console: 0 clicks, 15 impressions, avg position 1.8
  • Microsoft Copilot citations: 11 in 7 days
  • Amazon affiliate: 14 clicks (30 days), $0.00 earned, 0/3 qualifying sales
  • Top page: AI Investment Operating System — 124 views
  • New posts: 2 (AI Investment OS launch + Childhood Home AI Restoration)

Revenue: $0.00. Days running: 40. Copilot citations: 11. The tool beat the homepage.

Learning to Make Videos While Google Forgets I Exist

This week I built two tools and started something I’ve been avoiding for months: learning video.

The tools: Gaokao English Study App — an AI-powered English exam prep tool for China’s college entrance exam. And AI Chinese Story World — a tool that turns Chinese folktales and modern stories into short video scripts. The second one exists for a specific reason: I need to learn video. Text alone won’t be enough. The only way to learn is to build something that forces me to actually make one.

The numbers went backwards this week. Active users dropped slightly from 172 to 150. Google Search Console position slipped from 1.8 to 4.4 — worse, not better. Impressions dropped from 15 to 7. Zero clicks. Amazon Associates: 0 clicks. Ko-fi: 0 tips.

This is the part of building in public that’s easy to skip — the week you did real work and the metrics still went the wrong direction. So I’m writing it down exactly as it happened: position got worse, traffic dropped slightly, revenue stayed at zero. I shipped two tools anyway. Still here.

The one genuine win: I started learning video. Slowly, badly, with no audience. That’s where every creator who eventually mattered started.

This week’s numbers:

  • Google Analytics: 150 active users, 142 new users, ~1,500 events
  • Google Search Console: 0 clicks, 7 impressions, avg position 4.4 (was 1.8 last week)
  • Amazon affiliate: 0 clicks this week, $0.00 earned, 0/3 qualifying sales
  • Ko-fi tips: $0
  • New tools shipped: 2

Revenue: $0.00. Days running: 47. Two tools shipped. Learning video from scratch.

Bing Found Me Before Google Did — Two New Tools Shipped

This week I built two tools that don’t fit neatly into the “AI content tool” category I’ve been building: a Standup Timer that reminds you to get up from your desk before sitting kills you, and a Website Security Checker that scans any URL for basic security issues — HTTPS, headers, mixed content, open redirects. Both came from real frustration. I sit too long. I check security too manually. Now there are tools for both.

The unexpected story this week is Bing. Google Search Console is still quiet — 1 click from 5 impressions, average position 4. But Bing Webmaster Tools showed something different: 8 clicks from 200 impressions, 4% CTR. That’s more clicks in one week from Bing than I’ve gotten from Google in two weeks combined. The irony of building an English blog with AI and having Microsoft find it before Google is not lost on me.

GA4 this week (Aug 2–8): 188 active users (up from 150 last week, +25%), 179 new users, average engagement 39 seconds, 1,600 total events. Users are up but engagement time dropped — more people finding the site, fewer staying long. That gap between arriving and staying is the next problem to solve.

External backlinks: 360 total. The breakdown is honest: zfuye.org accounts for 342 of them (my other site), GitHub gives 14, and the rest — csdn.net, domainarrivals.com, indiehackers.com — contribute 1 each. The “360 backlinks” number looks impressive until you see the source. Most of it is self-linking. The 14 from GitHub and the handful from real external sites are the ones that actually matter.

Amazon Associates: $0.00. 0 clicks this month, 0 qualifying purchases. The 3-purchase requirement to stay in the program is still at 0/3. I’m not going to pretend the numbers look good — they don’t. I’m logging it because that’s what building in public means.

Ko-fi: 0 tips. Page views: 4. The tip button exists. No one has pressed it. That’s fine — I didn’t add it expecting tips. I added it so there’s a path for anyone who wants one.

This week’s numbers:

  • Google Analytics: 188 active users, 179 new users, avg engagement 39 sec, 1,600 events
  • Google Search Console: 1 click, 5 impressions, avg position 4, CTR 20%
  • Bing Webmaster Tools: 8 clicks, 200 impressions, 4% CTR
  • External backlinks: 360 (zfuye.org 342, GitHub 14, others 4)
  • Amazon affiliate: $0.00 earned, 0/3 qualifying sales
  • Ko-fi tips: $0
  • New tools shipped: 2 (Standup Timer + Website Security Checker)

Revenue: $0.00. Days running: 54. Bing clicked before Google did. Two tools shipped anyway.

📅 Days 54–62 — August 9–16, 2026

Hostinger Was Too Cheap to Ignore. So I Bought It. Then Had to Figure Out What to Build.

It started with a billing page. I was looking at my Hostinger account — I’d bought the original Premium plan two months earlier for around $33, covers 3 sites, runs until 2028 — and noticed the renewal price sitting there: $263 for another two years. Meanwhile, a brand new 4-year Premium plan was listed at $143.52. That’s a gap that’s impossible to ignore. Buy new, don’t renew old.

Then I remembered I’m enrolled in Hostinger’s own affiliate program. I used my referral link to buy the plan — REFERRALDISCOUNT -20%: −$28.70 — final price: $114.82 for 48 months, domain included (normally $13.99/year). And because of a 40% special promotion commission, I earned $45.93 in affiliate cashback, arriving September 28th. Net cost: $68.89 for 4 years of hosting plus a free domain. Buying your own affiliate product isn’t gaming the system — if you’re already recommending it to readers, you should be able to buy it through your own link.

Hostinger affiliate referral commission $45.93 earned — screenshot proof

My Hostinger affiliate dashboard — $45.93 commission earned from my own purchase. 40% special promotion applied.

If you want the same deal, here’s my Hostinger affiliate link — the 48-month Premium plan is genuinely the cheapest long-term hosting I’ve found, especially if you’re building multiple small sites. (Affiliate link — I earn a commission if you buy, at no extra cost to you.)

OK — so I had a new hosting plan. Now I needed a site to put on it.

My first instinct was to clone the model behind wheelpage.com — a spin-the-wheel decision tool. Zero maintenance once built. Pure SEO traffic. AdSense for monetization. I spent 30 minutes studying it: roughly 10,000 monthly visits, no paid ads, survives on “spin the wheel” and “random name picker” searches. The kind of site that earns money while you sleep. I asked Claude to run the AdSense math: at 10,000 monthly visits with US/UK traffic, AdSense pays $3–8 RPM — $30–$80/month at wheelpage’s current level. Scale to 100K visits and you’re looking at $300–$800/month. Not life-changing, but fully passive after the build.

Then I asked AI a harder question: given that I’m Chinese, and I already have an English blog going, what’s the actual best niche for a low-maintenance vertical site? The answer was Chinese fortune tools — zodiac calculators, BaZi charts, feng shui tools. Three reasons: (1) real English-language search volume from non-Chinese users curious about Chinese astrology, (2) most existing sites are either ugly or Chinese-only, leaving the English niche thin, (3) the content never goes stale. A person’s BaZi chart doesn’t change. The zodiac cycle doesn’t update. Build it once, leave it running.

I dropped the wheel idea. Registered chinesefortunetools.com on August 14, 2026. Built the first version — homepage, zodiac calculator, BaZi tool, guess-your-zodiac game — in 48 hours. Submitted to Google and Bing the same day it went live.

OMT Week 9 — 225 Users, 65 Copilot Citations, Google Still Quiet

GA4 (Aug 9–16): 225 active users (+20% from 188 last week), 218 new users, average engagement 1 minute 16 seconds — up from 39 seconds the week before. Games keep people on page longer than articles do.

Traffic sources: Direct/none: 195 users (me testing + dark social). GitHub: 9. Bing organic: 7. Zhihu: 4. Google organic: 6 sessions. Small numbers, but organic exists on both engines.

Google Search Console: 1 click, 8 impressions, avg position 14.6, CTR 12.5%. Indexing is the bigger problem: 204 pages indexed, 547 not. Worst buckets: 199 pages “Crawled — currently not indexed” and 242 “Discovered — currently not indexed.” Google sees them and skips them. Thin content on the programmatic pages is the most likely cause.

The week’s real surprise: Bing Copilot cited this blog 65 times in 7 days, averaging 3 cited pages per day. Not search clicks — AI citations. The blog is being pulled into Copilot answers without anyone clicking through. Whether that ever converts to actual traffic, I don’t know. But 65 citations at day 62 wasn’t something I expected.

Amazon Associates: $0.00. 0 clicks, 0/3 qualifying sales. Logged as always.

Ko-fi: $0. 4 page views. The button exists.

This week’s numbers:

  • Google Analytics: 225 active users, 218 new users, avg engagement 1m 16s, 1,900 events
  • Google Search Console: 1 click, 8 impressions, avg position 14.6, CTR 12.5%
  • Bing Copilot citations: 65 in 7 days, avg 3 pages/day
  • GSC indexing: 204 indexed, 547 not indexed
  • Amazon affiliate: $0.00, 0/3 qualifying sales
  • Ko-fi: $0

Revenue: $0.00. Days running: 62. Bing AI cited me 65 times. Google still hasn’t noticed.

ChineseFortuneTools: First 48 Hours

The new site went live on August 14 — two days before this entry. So these aren’t “week one” numbers. These are 48-hour numbers.

GA4 snapshot: ChineseFortuneTools homepage — 82 views, 30 active users. The guess-your-zodiac game got 50 views, 11 users. The Chinese zodiac calculator got 29 views. A zero-backlink brand new site already generating more page views than OMT’s own homepage (21 views) — in 48 hours. That’s either a strong signal or a reflection of how early-stage OMT still is. Probably both.

Submitted to Google Search Console and Bing Webmaster Tools on launch day. Nothing indexed yet — that’s normal. The real first-week numbers will show up in next week’s entry.

I didn’t start Chinese Fortune Tools because I had a business plan. I started it because I had some cheap server space and an itch to build something.

📅 Days 63–67 — August 17–23, 2026

The Claude tokens are simply not enough, so I also used DeepSeek, which cost me 50 RMB.

📅 Days 68-75 — August 17–23, 2026

I’ve bought two more domains to make the most of my Hostinger plan instead of leaving it underused — https://chinesenamecraft.com/ and https://chinarules101.com/.

📅 Days 76-84 — August 24–Sept 8, 2026

I’ve bought two more domains to make the most of my Hostinger plan instead of leaving it underused — https://hotpot101.com/.

📅 Days 85-97 — Sept 9–Sept 20, 2026

I’ve bought two more domains to make the most of my Hostinger plan instead of leaving it underused — https://chinatea101.com/ and https://cutdone.com. A lot has changed since my last update. What started as one simple blog has somehow turned into almost ten different little experiments. I built Hotpot101, started ChinaTea101, created CutDone, and even registered OnePersonCoHQ as a future home for all these projects. So far, I’ve spent around RMB 2300 in total. Revenue is still zero, and traffic is small. But I’m starting to understand something important: I build much faster than Google and real users can catch up. The biggest lesson this month isn’t how to build more. It’s how to stop building. A few months ago, my question was “Can I build websites with AI?” Now I know the answer is yes. The new question is much harder: Can any of these become a real small business?

Related Reading

One response to “I Built This Blog in 4 Days Using AI — And Now I’m Running a Live Experiment in Public (Follow Along)”

  1. I Bought Cheap Hosting Before I Had a Business Plan | OMT

    […] 🔴 Live Diary […]

    Reply

Leave a Reply

Your email address will not be published. Required fields are marked *