LearningAugust 17, 2026 · 7 min read

How to Learn a Language to Conversational Level in a Year

Half the country wants to do this. In Choice Mutual's bucket-list survey, 50.8% of people named learning a new language — second only to seeing specific landmarks in Europe. Almost nobody finishes, and the reason is not discipline. It is that nobody tells you the number. Learn a language to conversational level and you are signing up for something specific and measurable, and once you can see the figure the whole thing stops being mysterious and starts being a calendar. This guide gives you the number, where it comes from, and the four commitments that turn it into a year you actually complete.

A sunlit desk by a window with a stack of books, a closed notebook, a mug and headphones

What does "conversational" actually mean, and how many hours is it?

Conversational is B1 on the CEFR scale, and Cambridge puts B1 at 350 to 400 guided learning hours. Their published ladder runs 90–100 hours for A1, 180–200 for A2, 350–400 for B1, then 500–600 for B2 and 700–800 for C1. So the target is not vague at all. It is roughly an hour a day for a year.

B1 is the level where a language stops being a subject and starts being a tool. The official descriptor is dry — you can deal with most situations while travelling, describe experiences, give reasons and explanations — but what it feels like is more specific than that: you stop translating in your head and start following. You will still miss things. You will no longer be lost.

Write the number down, because it is the single most useful fact in this entire guide, and it immediately settles two arguments. An app that promises conversational in six months at fifteen minutes a day is offering you 91 hours, which does not even reach A2. And "I've been learning for years" usually means a few hundred scattered minutes, which is why so many people who have genuinely been at it for years are still at hello. The hours are the whole game.

Why do the numbers online disagree so wildly?

Because the two most-quoted sources are measuring different things. Cambridge's 350–400 is the hours to reach conversational. The famous Foreign Service Institute figures — 600–750 hours for Spanish or French, 2,200 for Japanese, Chinese, Korean or Arabic — are the hours to reach professional working proficiency, which is a much higher bar. Quoting one at the other is the most common mistake in this entire subject, and it is why the internet can tell you the same goal takes 350 hours or 2,200.

Three stacks of books at clearly different heights side by side, page edges facing the camera

The FSI has been training US diplomats since 1947, and their data is excellent — thousands of learners, structured programmes, measured outcomes. But read the fine print. The target is ILR level 3, "Speaking-3/Reading-3", which is roughly C1: able to work professionally in the language. The delivery is intensive full-time study, around 23 to 25 hours a week in class plus about 17 hours of self-study on top. And the students are diplomats selected in part for language aptitude. So a Category IV language at 2,200 hours means 88 weeks of essentially a full-time job, aiming at a level far beyond conversation.

None of which makes FSI wrong — it makes it inapplicable. If you are asking "how long until I can have a conversation," the FSI number is not your number. What FSI is genuinely good for is relative difficulty: whichever target you pick, Japanese will cost you something like three times what Spanish does. Use it to set expectations between languages, not to set your timeline.

What does an hour a day actually look like across a year?

An hour a day for a year is 365 hours, which lands inside the 350–400 band. That is the plan in one sentence. And the shape matters as much as the total: an hour daily is 365 hours, while a heroic three-hour weekend session twice a month is 72. Same feeling of effort, one-fifth of the result.

This is your first tracked commitment — the daily hour — and it is the one the research is most emphatic about. The advantage of spreading practice out rather than massing it is one of the most reliably replicated findings in learning science. A 2022 meta-analysis in Language Learning pooled 98 effect sizes from 48 experiments with over 3,400 learners and found a medium-to-large benefit for spaced practice in second-language learning specifically; classroom studies put the effect around d = 0.54. Notably, the gap widens the longer afterwards you measure. Cramming can get you through Friday's test. It cannot get you to B1.

Headphones resting on a train window seat with bright morning light outside

The practical question is where the hour comes from, and the answer is usually that it does not come from one place. A commute with headphones, twenty minutes of review at lunch, half an hour before bed — the research says the splitting is fine, and may even help. What is not fine is the month off. Missing a day costs you a day. Missing a month costs you the month plus everything that quietly faded while you were away, and that second part is what turns a one-year plan into a three-year one.

What actually goes in each level?

Three levels — A1, A2, B1 — and each one is a short list of things you can finish. A coursebook. A practice paper. One conversation. Not "learn 2,000 words", because nobody counts their own vocabulary and a target you cannot check is not a target.

A worn coursebook and a well-used notebook lying open-edged on a wooden table

This is your second tracked commitment — the ladder — and the design principle behind it is the same one that makes a half marathon training plan work: use the rungs the domain already has, and make every rung something with an end. Word counts are worth knowing but not worth tracking. Roughly 500 words at A1, 1,000–1,500 at A2, 2,000–3,000 at B1 — and note that the CEFR itself does not define vocabulary sizes at all. Those figures come from corpus research and the exam boards converging, not from the framework. They tell you how big a level is. They make terrible checkboxes.

So each level gets finishable objects instead. A1 is roughly the first hundred hours: learn the sounds properly and once, at the very start, because fixing pronunciation later is far harder than learning it right and it decides whether people understand you at all; finish one beginner coursebook end to end; pass an A1 practice paper; and introduce yourself out loud to a real human being. A2 is where past and future arrive, which is what lets you talk about anything other than right now — finish the coursebook, pass the paper, hold ten minutes about what you did yesterday, and read one news article written for learners without a dictionary. B1 is the coursebook, two practice papers in a row, thirty minutes with a native speaker without switching back, and one episode watched with subtitles in the language rather than in yours. Pick something everyday for that last one; period drama and medical drama run on vocabulary from another planet.

Why book an exam if you don't need the certificate?

Because a date you cannot move is the difference between learning a language and meaning to. Nobody needs the certificate. Everybody needs the deadline. Register early, and register on purpose — language exams take entries months ahead and close them without ceremony, so missing the window usually means waiting for the next sitting rather than paying a late fee.

An empty exam hall with rows of identical desks, a pencil and eraser on the nearest one

This is your third tracked commitment — exam day — sitting about eleven months out with four checkpoints in front of it: pick the language and the date, wrap up A1, wrap up A2, then register. That third-to-last one carries the whole plan. Without a fixed date this quietly becomes a hobby you drift out of, and the drift is invisible while it is happening.

One decision belongs at the very start and is worth being blunt about: pick one language. Splitting across two is the single most common way this ends, because neither one ever reaches the level where it starts being fun, and "fun" is not a soft consideration here — it is the mechanism that gets you through months eight to twelve. If you genuinely do not want a certificate, substitute a date that is equally immovable: a trip booked, or a conversation you have promised someone.

Will passing the exam mean you can actually speak?

That depends entirely on which exam, and it is worth checking before you build a year around it. The big East Asian tests largely do not assess speaking: the JLPT is entirely multiple-choice with no speaking or writing section at all, the HSK splits speaking into a separate exam called the HSKK, and TOPIK runs its speaking test separately from the main paper. European exams are the opposite — Cambridge, DELF and DELE all include a live speaking paper as a full component.

So there are two different traps and you should know which one you are walking into. If your exam has no speaking section, nothing in the entire year will force you to open your mouth, and you can arrive at the certificate genuinely unable to hold the conversation it appears to certify. If your exam does have one, you get tested exactly once, at the very end, which is far too late to discover that you have never done it.

Either way the fix is the same, and it is why every level in the ladder above carries one check you can only pass by speaking: introduce yourself out loud at A1, ten minutes about yesterday at A2, thirty unbroken minutes at B1. They are deliberately small and deliberately early. The habit worth building at A2 — and the one most people skip — is speaking with someone once a week. It is not a gate. Skipping it is simply how the certificate and the conversation come apart.

What is worth keeping from the year?

Four things, and only one of them is the certificate: the first whole exchange you got through with a native speaker, the first thing you finished without subtitles in your own language, the exam room on the day, and the certificate itself.

A small cafe table by a window with two cups facing each other and one chair pushed back

This is the fourth and last tracked commitment, and it is here because a year is long enough that you will forget what the beginning felt like. The first real exchange is the one that matters most and the one nobody thinks to record — it is usually clumsy, roughly four minutes long, and the moment the whole thing stops being theoretical. The certificate is the least interesting of the four, which is exactly the point: results appear online first and the paper takes weeks longer, by which time you have already had the conversations that were the actual goal.

The number is 350 to 400 hours. The plan is an hour a day, a ladder of three levels made of things you can finish, a date you cannot move, and one check per level that you can only pass by talking. When you are ready to start counting, set it up as a bucket-list goal in DreamPaths and let the year keep score. The hours are the whole game — and an hour a day is a very ordinary thing to ask of yourself.

Frequently asked questions

How many hours does it take to become conversational in a language?
Around 350 to 400 guided learning hours, which is Cambridge's published figure for B1 — the level most people mean by "conversational". Their full ladder is 90–100 hours for A1, 180–200 for A2, 350–400 for B1, 500–600 for B2 and 700–800 for C1. An hour a day for a year is 365 hours, which lands squarely inside the B1 band. Fifteen minutes a day for a year is 91 hours, which does not reach A2.
Why do people say Japanese takes 2,200 hours?
That figure comes from the US Foreign Service Institute, and it measures something different. FSI hours are for professional working proficiency — ILR level 3, roughly C1 — delivered as intensive full-time study of about 23–25 classroom hours a week plus around 17 hours of self-study, to diplomats partly selected for language aptitude. It is not the number for holding a conversation. What FSI is genuinely useful for is relative difficulty between languages: whatever your target, Japanese costs roughly three times what Spanish does.
Is fifteen minutes a day on an app enough?
Not on its own, and the arithmetic is the honest answer rather than a criticism of apps. Fifteen minutes a day is 91 hours a year, which is short of even A2 at 180–200 hours. Apps are genuinely good at vocabulary and at the daily-habit problem, which is most of the battle. They just cannot compress the total. If fifteen minutes is what your life allows, the honest expectation is A2 in about two years — still real progress, just not the year-to-conversational plan.
Do I have to take an exam?
No, but you do need a date you cannot move, and an exam is the cleanest one available. A booked trip or a conversation you have promised someone works just as well. If you do choose an exam, look up the registration closing date the same day you choose it — language exams take entries months ahead and close them without a late-fee option, so missing the window usually means waiting for the next sitting.
Will passing the exam mean I can actually speak?
It depends which exam, and it is worth knowing before you build a year around one. The JLPT is entirely multiple-choice with no speaking or writing section; the HSK splits speaking into a separate exam (the HSKK); TOPIK runs its speaking test separately. Cambridge, DELF and DELE all include a live speaking paper. So either nothing in your year forces you to speak, or you get tested exactly once at the very end — which is why it is worth putting one speaking check into every level rather than leaving it to the exam.

Turn this guide into your plan

DreamPaths tracks bucket-list goals as habits, deadlines, budgets and milestones — the exact structure this guide is built on.