Private English Tutor vs Preply or italki: Costs, Rules and What You Actually Get
Marketplaces are the best place to find a tutor and often the worst place to keep one. A tutor who has worked both ways explains the trade-offs.
Shadowing means listening to a short piece of native speech and repeating it out loud at the same time, copying the rhythm, stress and melody, not just the words. Done correctly for 15 minutes a day, it changes how you sound faster than any other single exercise I know.
The short answer
The shadowing technique for English works like this: choose 20 to 40 seconds of clear American speech with a transcript. Pass one: listen twice and mark the stressed words. Pass two: read the transcript aloud together with the audio, half a beat behind the speaker. Pass three: shadow without the text, then record yourself and compare. Repeat the same clip for three to five days before moving on. The mistakes that make it useless are shadowing too fast, copying words instead of rhythm, and never recording yourself.
Why the shadowing technique works when "listen and repeat" does not
Most learners have done "listen and repeat" for years. You hear a sentence, the audio stops, you say it. It builds vocabulary and it does almost nothing for accent, because you have time to translate the sentence back into the sound system of your first language before you speak. Shadowing removes that time. You speak while the speaker is still talking, so you are forced to copy the pitch going up on really, the shrinking of to into tuh, the way want to becomes wanna. You do not have time to think, and that is the point.
It also trains your ear. In the first week of shadowing almost every student says the same thing: "I did not know she said it like that." The stressed words, the dropped sounds and the links between words are invisible until you try to copy them in real time. That is why I put shadowing in almost every lesson plan, and why my short YouTube videos are built as shadowing material: one sound or one phrase, said slowly and then at normal speed.
Choosing audio: short, clear, with a transcript
The clip matters more than the effort. Four rules:
- 20 to 40 seconds. Longer clips turn into passive listening. You want something you can repeat ten times in a session without losing focus.
- One speaker, clear audio, General American. Interviews and podcasts with two people talking over each other are bad material. News anchors, TED-style talks, audiobook samples and my own videos work well. If you are aiming for a neutral American accent, do not mix in British material for now, because consistency is what makes an accent sound natural.
- A transcript you trust. Auto-generated captions are good enough for pass one, but check them. A wrong word in the transcript will teach you a wrong word.
- Slightly above your comfort level. If you can understand it fully on the first listen, it is too easy to teach you anything about rhythm. If you understand less than half of it, it is too hard to shadow.
Stay away from sitcoms and movies at the start. They are full of overlapping speech, background music and character voices. Come back to them once you can shadow a news clip comfortably.
The three passes, minute by minute
This is the exact routine I give students, and it fits in 15 minutes.
| Pass | Time | What you do | What you are training |
|---|---|---|---|
| 1. Listen and mark | 3 min | Listen twice with the transcript. Underline the two or three loudest words in each sentence. Draw a line between words that link. | Your ear: stress, linking, reductions. |
| 2. Shadow with text | 5 min | Play the clip and read aloud with it, staying half a beat behind. Repeat four or five times. Match the melody first, sounds second. | Rhythm and intonation. |
| 3. Shadow blind, then record | 7 min | Put the transcript away. Shadow twice. Then record yourself saying the clip alone, and play your recording and the original back to back. | Automatic production and self-correction. |
Pass one: listen and mark
Take a sentence like "I actually think we should wait until Friday." A native speaker does not say seven equally loud words. She says i ACtually THINK we should WAIT until FRIday. The words in capitals get the time and the pitch. Everything else gets squeezed. Mark those stressed words on the transcript before you speak, because if you do not know where the stress is, you will spread it evenly, and even stress is the single biggest reason speech sounds "foreign" even when every sound is right.
Pass two: shadow with the text
Start the audio and speak with it, staying just behind the speaker, like an echo. Do not try to be perfect. Try to be on time. If you fall behind, drop a word and catch up rather than stopping. In this pass you are copying the shape of the sentence: where the voice rises, where it falls, where it speeds up over we should and slows down on wait.
Pass three: shadow blind, then record
Now close the transcript and shadow twice from memory. Then record yourself saying the whole clip without the audio. Listen to your recording and the original back to back and ask three questions: Are my loud words the same as hers? Did I link wait until into way-tuntil? Did my voice fall at the end of the statement? Write down one thing to fix tomorrow. One, not five.
What to listen for: stress, linking and melody, in that order
Students usually start by listening for individual sounds, which is the wrong order. Here is my priority list.
1. Sentence stress. Which words carry the meaning? Usually nouns, main verbs, adjectives and negatives. Can't is stressed, can is not, and that difference is what Americans actually listen for. Grammar words like to, of, and, a shrink to a schwa: tuh, uhv, uhn, uh.
2. Linking. American English joins words together. A consonant at the end of one word attaches to a vowel at the start of the next: an apple becomes a-napple, turn it off becomes tur-ni-toff. A T between vowels becomes a soft flap: get it sounds like geddit. If this is new to you, read my guide to connected speech and linking before your next session, because shadowing is where linking finally becomes physical.
3. Melody. Statements fall at the end. Yes/no questions rise. Lists go up, up, up, down. Copy the pitch movement even if it feels exaggerated. When students record themselves, their pitch movement is always flatter than they think. Overdo it on purpose and the recording will sound normal.
4. Individual sounds, last. Once the rhythm is right, listen for one target sound per clip: every TH, every R, the vowel in can't. Trying to fix everything at once fixes nothing.
The four mistakes that make shadowing useless
I see the same errors in nearly every new student who has "tried shadowing and it did not work."
- Shadowing too fast. If the clip is at full speed and you are behind by a whole sentence, you are not shadowing, you are mumbling. Slow the audio to 0.75 on YouTube for the first two days, then go back to normal speed. Speed is the last thing to add, not the first.
- Copying words instead of rhythm. Getting all the words out on time with flat stress is a memory exercise, not an accent exercise. If your version has no loud and quiet words, start pass one again.
- Never recording yourself. Without the recording, you believe you sound like the speaker. You do not, and neither do I when I shadow German. The recording is where the learning happens. Two minutes of honest comparison is worth twenty minutes of shadowing.
- Changing the clip every day. One clip, three to five days. The first day you learn what the speaker does. The third day your mouth starts doing it. The fifth day it is automatic, and you can move on. New clip every day means you are always on day one.
There is a fifth problem that is not really a mistake: you cannot hear certain errors in your own recording. A Korean speaker often cannot hear the difference between their R and the speaker's. A French speaker hears an H that is not there. This is the part shadowing alone cannot fix, and it is the first thing I listen for when a student shadows in a lesson. I stop the audio, show the mouth position on camera, and we repeat the phrase until it is right. Then the home practice works, because you are finally practicing the correct thing.
A 15-minute daily shadowing routine for a busy week
Here is how the three passes spread across a week with one clip. Every session is 15 minutes; if you only have ten, cut pass two, never pass three.
| Day | Focus | Speed |
|---|---|---|
| Monday | Passes 1 and 2 only. Mark stress and links. Get on time. | 0.75 |
| Tuesday | All three passes. First recording. Note one fix. | 0.75 then 1.0 |
| Wednesday | All three passes. Focus on the fix from Tuesday. | 1.0 |
| Thursday | Pass 3 only, three times. Add one target sound (for example every TH). | 1.0 |
| Friday | Final recording. Compare with Tuesday's. Say the clip to someone, or in a voice message. | 1.0 |
On the weekend, rest or pick the next clip. Over a month you will have shadowed four clips deeply, which is far more useful than thirty clips once. To see how this fits with reading aloud and narrating, see how to practice English speaking alone.
What shadowing cannot do, and what to pair it with
Shadowing builds rhythm, melody and speed. It does not build conversation skills, because you are copying, not creating. Students who only shadow can sound very American while reading and go straight back to their old accent when they have to find the words themselves. The fix is to shadow a clip and then talk about it for two minutes in your own words, keeping the same stress and linking. That transfer step is what I do with students at the end of most lessons.
Pair it with active listening, too. If you cannot hear the reductions in fast speech, you cannot copy them, and listening practice with transcripts is the shortest way to fix that.
Finally, shadowing works much better when someone checks your recordings once a week and catches the sound you are practicing wrong before it becomes a habit. That is roughly what the one-lesson-a-week plan is for. If nobody has ever listened closely to your speech, book a first lesson, shadow a clip with me, and you will know within ten minutes which sounds your recording is hiding from you.
What to do this week
Pick one clip today: 30 seconds of a news anchor, a clear TED speaker, or one of my videos on YouTube. Copy the transcript into your notes and mark the stressed words in each sentence. Tomorrow, do all three passes at 0.75 speed and save your first recording with the date in the file name. Do the same clip every day until Friday. On Friday, listen to Tuesday's and Friday's recordings back to back. You will hear the difference, and that is what will keep you going into week two.
Questions students ask
Is the shadowing technique for English useful at B2 or is it only for beginners?
It is most useful at B2 and above. At that level your grammar and vocabulary are solid, and what holds you back is rhythm, linking and stress, which are exactly what shadowing trains. Beginners struggle with it because they cannot follow the meaning fast enough.
How long should I shadow each day?
Fifteen minutes is enough if you follow all three passes and record yourself. Thirty minutes without recording is less useful than ten minutes with it. Daily short sessions beat one long session at the weekend.
Should I shadow with or without the transcript?
Both, in that order. Use the transcript for the first pass so you can mark stress and linking, and for the second pass so you stay on time. Then put it away for the third pass, because reading and speaking use different parts of your attention and you need the sounds to work without the text.
Can I use movies and TV shows for shadowing?
Later, yes. Start with single-speaker material like news, talks and short pronunciation videos, because the audio is clean and the speech is consistent. Once you can shadow those at full speed, a slow-talking character in a drama is a good next step. Avoid comedies with overlapping speech.
How do I know if my shadowing is actually improving my accent?
Compare recordings a week apart and listen for three things: stressed words in the right places, words linked together instead of separated, and pitch that falls at the end of statements. If none of those has changed, slow the audio down and repeat pass one. If you are unsure what you are hearing, that is what a lesson is for.
