From Zero to HSK: My Notion-Based Chinese Learning Journey
The honest story of going from no Chinese to passing HSK levels, and the Notion system that carried me through the plateaus, the burnout, and the rebuilds.
I started Chinese the way most people do: an app, a burst of motivation, and a vocabulary that evaporated the moment I closed the tab. Two years later I'd passed HSK levels I once thought were for other people. The thing that changed wasn't a course or a tutor — it was building a system in Notion that made the work survivable. This is the honest version of how that went, plateaus and dead ends included.
Month 1: The shiny-app phase
For the first few weeks I bounced between flashcard apps. They were fine for raw memorization, but everything lived in walled gardens. The word I learned from a podcast couldn't talk to the grammar point that explained it, and I had no record of why I'd added anything. I was accumulating cards, not understanding.
The turning point was small. I started keeping a single Notion page — just a running list of words with the sentence I'd heard them in. That context, the sentence, turned out to matter more than the word.
Month 2–3: From a list to a database
The flat list got unmanageable around 200 words, so I converted it to a database. This is the moment Notion started pulling its weight. Each word got a Status, a Source, an HSK level, and an example sentence. Suddenly I could filter: show me only HSK 2 words, only the ones still marked New, only words I'd mined from real listening.
The lesson here, which took me embarrassingly long to learn: a list tells you what you have; a database tells you what to do next. That difference is the whole game.
The first plateau (and the over-engineering trap)
Around month four I hit the wall everyone hits. Progress felt invisible, and I responded the way a lot of builders do — I rebuilt the system. I added a points engine, color-coded tone tags, embedded audio, and a relation between every conceivable thing. It was beautiful. I stopped using it within two weeks.
Why? Because every extra field was a tiny tax on adding a word, and the friction killed the habit. The system you maintain beats the system you admire. I tore most of it out and went back to four fields. My streak recovered almost immediately.
Month 5–8: Making review automatic
The breakthrough that mattered was a Due Today view. I stopped deciding what to review — a simple date formula decided for me. New words came back the next day, Learning words after three, Known words after two weeks. I'd open one view, work the list, stamp each row with today's date, and the reviewed words dropped off as I went.
This sounds trivial. It was the single highest-leverage change I made. Decision fatigue, not knowledge, is what ends most study streaks. Removing the daily "what should I study" question removed the off-ramp.
Connecting grammar to words
Reading didn't click until I related my Grammar database to my Vocabulary database. When I learned a resultative complement, I'd link the words that used it. Opening a grammar entry showed every word it touched; opening a word showed the pattern it lived inside. Grammar stopped being abstract rules and became the glue between things I already knew.
The HSK push
When I committed to an HSK exam, I added one view: a gallery grouped by HSK level showing the gap between what the level required and what I had at Known status. Watching each level fill in was the motivation that carried me through the boring final stretch. I didn't need a streak app — the visible gap closing was enough.
For the exam itself I added a small Mock Tests database, logging each practice score and the specific question types I bombed. That turned vague anxiety into a concrete weak-points list I could drill.
What I'd tell someone starting today
- Start with a list, not a system. Don't build the elaborate workspace on day one — you don't yet know what you need. Let the pain points tell you what to add.
- Capture the sentence, not just the word. Context is what makes vocabulary stick and what makes your database worth rereading.
- Automate the review decision. A crude date formula beats a perfect SRS you'll abandon. The goal is to never decide what to study.
- Cut anything that adds friction to capture. If a field makes adding a word slower, it had better earn its keep. Most don't.
- Make progress visible. A rollup of Known words or a level-by-level gallery does more for motivation than any gamification layer.
The system that survived
What I use now is almost embarrassingly simple: Vocabulary, Grammar, and Characters databases; a few relations; one date formula; a Due Today view; and a dashboard of linked views. No points, no badges, no audio embeds. It survived two years and an HSK exam precisely because it's boring.
The Notion setup didn't teach me Chinese — sentences, listening, and a lot of repetition did that. But it kept me showing up on the days I didn't want to, and over two years that's the only variable that actually mattered.