The best way to learn a language before an exam is a daily 90-minute loop with three legs: graded input tied to test topics, timed writing under a clock, and out-loud speaking practice. Do all three for 6 to 8 weeks and match the exact task types on your test. Skip any leg and scores flatten fast.
Your tutor today
The question this case study really answers
Is it possible to bridge an entire CEFR level within 8 weeks while keeping your job, staying in touch with friends, and still liking the language you’re learning once it’s over? Every exam candidate is silently wondering that, and it’s the truthful intent lurking behind searches for “best way to learn a language” once the test date is locked in.
What follows traces one candidate’s 8-week push toward TOPIK II Level 3, day by day, error by error. It’s a case study rather than a magic formula. Still, the underlying pattern carries over to IELTS, TOEFL, EIKEN, JLPT, DELF, or any structured language exam standing between you and a visa, a scholarship, or a job.
Here’s the short answer, stated early so it’s easy to lift: the most effective way to prepare a language for a test is a daily 90-minute routine resting on three legs. Graded input mapped to exam topics. Timed writing against a genuine clock. Speaking aloud with feedback. Drop any single leg, and the score curves go flat by week 4.
The setup: 8 weeks, TOPIK II, aiming at Level 3
Introduce the candidate. Sung, 19, is a first-year university student dividing her time between Seoul and the town she’s from. She’s been studying Korean for about two years through classroom courses plus a Netflix habit. Her scholarship renewal hinges on TOPIK II Level 3, which sits at the mid-B1 boundary. She has 56 days.
The starting whiteboard looked like this:
- Current level: a candid self-test placed her near low B1 in reading, high A2 in writing, A2 in speaking.
- Target: TOPIK II Level 3 (requires 120+ out of 300 across reading, writing, listening).
- Weekly time budget: 10 hours maximum. She also holds a part-time job.
- Tools already loaded: Anki, a bilingual TOPIK prep book, a single language app, YouTube.
- The actual problem: her Korean felt comfortable during flashcard recall and frantic inside a timed writing box. The exam demanded the reverse.
That final bullet is the case in a nutshell. Most exam candidates aren’t underprepared. They’ve prepared for the wrong task shape.
Week 0: the diagnostic that spoke honestly
Before any planning, Sung completed a full past-paper sitting under genuine exam conditions. Kitchen timer on the desk, no pausing, no dictionary. The outcome was brutal in a helpful way:
- Listening: 62/100. Solid on short dialogues, collapsed during the 4-minute lecture.
- Reading: 71/100. Confident with short passages, ran out of clock on the longer ones.
- Writing: 34/100. Abandoned the 600-character essay half-finished.
- Speaking (self-recorded mock): a hesitation every 3 to 4 seconds, plus heavy L1 grammar transfer.
This kind of diagnostic isn’t motivational. It’s architectural. It reveals where the 90 minutes truly belong. Sung’s figures pointed to writing first, then speaking, then listening endurance, and reading pace last.
Without a diagnostic, you’ll over-study whatever section you’re already strongest in. That’s the single most frequent exam-prep error, whatever the language.
Most exam candidates aren't underprepared. They've prepared for the wrong shape of task.
Praktika
The daily routine: 90 minutes across three blocks
What emerged as a plan wasn’t clever. It was dull, and that dullness is why it succeeded. Six days weekly, one rest day, 90 minutes split into three unequal blocks.
Block 1 (30 minutes): graded input mapped to an exam topic. Either one graded podcast episode or one past-paper reading passage on a TOPIK-style theme (health, education, technology, environment). She read or listened once at normal speed, once again with the transcript, then logged 5 fresh vocabulary entries into Anki along with a complete example sentence, never a bare gloss.
Block 2 (30 minutes): timed writing. A single prompt drawn from the TOPIK writing pool. Clock running. Daily. She began with 200-character short answers in week 1 and built up to the complete 600-character essay by week 4. The key wasn’t producing more text. It was composing against the clock from the very start, letting the panic turn into muscle memory.
Block 3 (30 minutes): speaking out loud. This is the block most self-learners abandon, since doing it alone feels awkward. Sung didn’t abandon it. Five days weekly she ran a spoken practice session with an AI tutor, using that day’s reading topics as conversation openers, then went through the pronunciation feedback together with one grammar correction.
That’s the entire routine. No 4-hour marathons. No cramming. No YouTube study playlists pretending to count as study. Ninety minutes, three blocks, six days.
Week by week: what genuinely moved
The case-study log spanned all 8 weeks. Below is the condensed version, showing what changed plus where the plan nearly collapsed.
Weeks 1 to 2: the ugly groundwork
Writing scores remained weak. Listening back to the speaking recordings was painful. Reading felt slower rather than quicker, because Sung had quit skimming and begun truly parsing grammar patterns. This is normal. Any exam plan’s opening fortnight looks like regression on paper. Trust the routine.
The lone measurable victory by day 14: vocabulary retention. Since every fresh word entered Anki attached to a genuine sentence from a genuine passage, recall hit roughly 85% at next-day review, against about 40% for the bare-word cards she’d relied on before.
Weeks 3 to 4: the writing breakthrough
Around day 22. something gave way. The 600-character essay stopped seeming impossible. Sung had absorbed a 4-paragraph template (introduction, point A with an example, point B with an example, short conclusion) and could now devote her writing block to content instead of structure.
Her mock writing score leapt from 34 to 51 by week 4.’s end. That’s the largest single-week gain recorded in the whole study. Not because she suddenly possessed more Korean, but because she quit burning the opening 8 minutes of every writing block fretting over how to begin.
The biggest score jump in the case study didn't come from knowing more Korean. It came from no longer burning the opening 8 minutes of every writing block fretting over how to begin.
Praktika
Weeks 5 to 6: the speaking compounding
This is the stretch where the daily speaking block earned its keep. Sung’s writing began reading more naturally, carrying fewer literal L1 translations, because she’d spent a month voicing those same grammar patterns aloud. Speaking practice is no luxury add-on for a written exam. It locks grammar in through a route silent study can’t reach.
Listening stamina grew too. The 4-minute lectures that had beaten her in week 0 now landed around 78% comprehension on the first pass. Not because her ears quickened, but because her working memory was freer, no longer busy translating every single phrase.
Weeks 7 to 8: taper and past papers
The final 14 days changed shape. Sung cut new input almost completely and completed full past papers under exam conditions three times weekly, reviewing only the mistakes afterward. It’s the exam-prep version of a marathon taper: reduced volume, sharpened specificity, sleep protected.
One habit she maintained: still 20 minutes of daily speaking, right through the taper. Fluency fades quickly once your mouth goes unused.
The outcome: what the score card showed
Sung took TOPIK II on day 57. Her composite result came in at 148/300, clearing Level 3 comfortably (120+ required). Section by section:
- Listening: 82 (from 62)
- Reading: 79 (from 71)
- Writing: 58 (from 34)
Writing posted the largest jump, aligning with where the plan concentrated effort. Reading showed the smallest, since it was already the strongest leg. That’s how a properly diagnosed plan appears on a score card: your weakest section gains the most.
One candid caveat: her speaking never underwent formal testing (the standard TOPIK II track includes no speaking section). Yet her self-reported confidence in Korean conversation doubled, which is precisely why the writing scores also climbed. The two can’t be pulled apart.
5 portable lessons for any language exam
Remove the specifics, and this is what genuinely carries across to any language, any exam, any 8-week window.
Lesson 1: diagnose before planning
Complete a full past paper under genuine conditions during week 0. Don’t skim your weakest section. Use a timer. Score yourself honestly. Every hour sunk into the wrong section is an hour gone forever.
Lesson 2: learn the test’s shape, not merely the language
IELTS Writing Task 2, TOPIK’s 600-character essay, DELF’s production orale, JLPT’s reading comprehension: each carries its own shape. Master that shape. Templates aren’t cheating. They’re scaffolding that releases your working memory to focus on content.
We’ve written a companion article about this for a European exam, covering it in greater depth: the DELF B2 preparation honest FAQ walks through the mechanism.
Lesson 3: write against the clock from day one
Never “work up” to timed writing. Begin timed on day one, even if your output is only two sentences. Handling time pressure is a distinct skill from producing language. Train both in parallel, never sequentially.
Lesson 4: speak aloud daily, even for a silent exam
Speaking practice is what converts passive grammar into active grammar. It’s why your writing starts reading like sentences instead of translations. Twenty minutes a day is the minimum. If your exam includes a speaking section, make it thirty.
This is the block self-learners sacrifice first, since solo practice is uncomfortable. It’s also the block dividing a Level 3 candidate from a Level 4 one.
Speaking practice is what converts passive grammar into active grammar. It's why your writing starts reading like sentences instead of translations.
Praktika
Lesson 5: taper, never cram
The final 10 to 14 days belong to past papers and review, not fresh material. Vocabulary added during the last week probably won’t be retrievable on exam day. Guard your sleep. A rested brain outscores a crammed one.
Where the tools fit: a quick, honest comparison
Sung rotated four tools. Below is what each genuinely contributed, with honest drawbacks.
| Tool | Strengths | Weaknesses | Approximate cost |
|---|---|---|---|
| TOPIK prep book (bilingual) | Authentic past papers, official rubric, drills by section | No speaking practice at all, no feedback loop, static material | $25 one-off |
| Anki (with real-sentence cards) | Spaced recall, vocabulary retention | Only as strong as the cards you make; dull on your own | Free |
| Praktika | Daily spoken sessions with an AI tutor, instant grammar and pronunciation feedback, topics adapt to you | Not a TOPIK-specific curriculum; the topic agenda is yours to drive | ~$8/month |
| YouTube (graded Korean channels) | Complimentary listening input, cultural background | Zero feedback, zero accountability, drifts easily into passive viewing | Free |
No tool handled the whole job on its own. A book by itself wouldn’t have exercised her mouth. A speaking app by itself wouldn’t have drilled the exact essay template. Anki by itself would have taught her nothing beyond recognising isolated words.
The combination made it work. If your budget covers just one paid tool, this case study points toward spending it on the block you’re most prone to skipping solo — for most exam candidates, that’s daily spoken practice. Anki, the book, and YouTube are free or close to it.
For a wider view of what justifies a language app’s subscription in an exam setting, we unpacked that in the 5 mechanisms behind real progress.
What this case study is not
A few candid limits, since case studies get over-generalised rapidly.
- This was a single candidate, a single exam, a single 8-week window. Your mileage will differ based on your L1, your starting level, and the hours you can genuinely protect.
- Sung already carried two years of Korean entering day one. Beginning from zero? Then 8 to B1 isn’t realistic. Sixteen to twenty-four weeks is the honest figure.
- The 90-minute daily routine assumes 90 minutes protected minutes. If your ceiling is 45, halve every block and anticipate roughly half the gains.
- TOPIK II doesn’t test speaking directly. For TOEFL or IELTS, layer a dedicated speaking-section drill on top of the 30-minute out-loud block.
Want the deep-dive on why apps are built the way they are for exam prep? This piece on choosing a language app for a specific outcome lays out the criteria in greater detail.
The self-check
A single question — answer it immediately, aloud if you prefer.
Suppose you sat your target exam this Sunday under genuine conditions: which section would come out lowest, and is that section actually getting the bulk of your study time?
If the answer is no, your plan is staring at you. Shift the hours. Diagnose during week 0, assemble the 90-minute routine, guard the speaking block even for a silent exam, and taper through the final fortnight.
And if the block you keep dodging is the out-loud one, that’s the one to sort out this week. You can start a free conversation with a Praktika tutor tonight — in Korean or any other exam language — and hear the distance between what you can write and what you can say. That distance is this case study’s entire point.