Three methods. The same hours. Very different destinations.

Each of these works. People have reached fluency with all three. What separates them is how quickly you see progress, how high you can eventually get, what they take out of you, and how likely you are to stop.

Read the one that matches your target. Then read the one below it, so you know what you are giving up.

1

Expression‑first

Target: A1, fast · typical run 4–12 weeks

Speak from the first day. Learn words deliberately, learn the grammar you need to build sentences, and get a tutor early. This is what almost every school, app and language course sells, and for its actual purpose it is the correct choice.

What you actually do

Week 1The first few hundred words, chosen by frequency. Pronunciation of the alphabet and the sounds that do not exist in your language. Your first tutor session before you feel ready.
Weeks 2–4Present tense, the verbs you need to say what you want, and sentence patterns you can reuse. Speaking two or three times a week. Listening in the background, but it is not the priority.
Weeks 5 onwardMore situations, more phrases, more speaking hours. Past tense when you need it. Repetition of the same conversations until they run without thinking.
What it costs
The most effort per hour of any method here. Conversation while you are bad at a language is genuinely tiring in a way that watching and reading is not, and that tiredness is the main reason people stop.
How you know it is working
Fast, obvious, external. You will have your first real exchange within weeks, which is exactly why this method feels so good and gets recommended so widely.
Where it stops
It gets you to basic conversation quickly and then flattens. Comprehension lags a long way behind production, which is a strange and frustrating place to be: you can say more than you can understand.
Who should not use it
Anyone who wants to end up genuinely good. Leaving this plateau later costs more than starting differently would have cost at the beginning.
2

The combination

Target: B1 · typical run 6–12 months · the right choice for most people

Lead with listening and reading, add writing and grammar in the middle, and bring speaking in last with the intensity rising towards your deadline. It is the compromise, and compromises are underrated: this one gets most of the ceiling of the slow method at a fraction of its psychological cost.

What you actually do

Phase 1 — roughly the first thirdListening and reading, most days, plus deliberate vocabulary. No speaking, little grammar. The goal is a large passive vocabulary and a feel for how the language sounds.
Phase 2 — the middle thirdKeep the input going. Add grammar properly — it will make sense far faster now, because you have heard the patterns before anyone named them. Start writing: same production as speaking, but with time to think and check.
Phase 3 — the last thirdSpeaking, starting gently and increasing towards the deadline. The input never stops. Most people find this phase moves much faster than they expected, because they are assembling something already there rather than building from nothing.
What it costs
A few months where progress is real but hard to show anyone. Long enough to require some patience, short enough that most people do not quit.
How you know it is working
You start catching words, then phrases, then whole sentences in things you are watching. It creeps up rather than arriving.
Where it stops
Comfortably conversational, with a foundation solid enough to keep building on if you later decide you want more.
Who should not use it
People with a hard deadline inside three months, and people who genuinely want a very high level and have the patience for method three.
3

Input‑first

Target: B2 and above · typical run 12–24 months

Months of listening and reading before you produce anything at all. Output comes late, and when it comes it comes quickly, because you are giving voice to a language you already understand rather than assembling one you do not.

What you actually do

The long phaseListening and reading, daily, for months. Weighted towards listening early on, because that builds the sound system that your accent will later depend on. Optional light vocabulary work. No speaking.
The switchWriting first, then speaking, once you can follow real material comfortably. This is not a hard date — it is a readiness condition.
After the switchSpeaking rises fast. The input does not stop and never should. Most of the remaining work is turning what you understand into what you can produce.
What it costs
This is the honest part that most people advocating this method skip. There will be months where you cannot demonstrate anything to anyone, including yourself. Friends will ask you to say something and you will not be able to. There is no external validation available for a long time, and that, not difficulty, is what makes people quit.
How you know it is working
Only through comprehension. Material that was noise becomes words. If you need any other signal than that, this is the wrong method for you.
Where it stops
Highest ceiling of the three, particularly for pronunciation — which is the one area where doing things in this order has the strongest evidence behind it.
Who should not use it
Most people. Not because it does not work, but because needing visible progress is normal, and a method you abandon at month four beats nothing but loses to method two.

Where I have to be careful. This is the method I used for French, and I delayed speaking for around nine months. I want to be straight about what that is worth: it is my experience, not evidence. There is no controlled study showing that a nine‑month delay produces a better outcome than an input‑heavy schedule that includes some speaking.

What is well supported is the narrower claim underneath it: you cannot reliably produce a sound you cannot yet hear, and training perception measurably improves production. That is why the order matters for accent. For grammar, the case is much weaker, and I say so.

Not sure which one is yours?

It comes down to your target, your deadline, the language, and how you react to months of invisible progress. Five questions.

Find your method