A teaching certificate is a filter, not a price. It gets your application read by schools and it satisfies the box on a platform's profile, and once you are through that gate it stops doing anything at all: nobody has ever paid a teacher more per hour because the certificate was the longer one. What sets the rate is a different question entirely, and it is who found the student.
So the decision that matters is made early and quietly. Every hour you teach is either building somebody else's business or building the list of people who would follow you, and the two feel identical while you are teaching them.
The six this cluster hangs on
- talk less than the students. Teacher talking time is the single most common problem in an observed lesson, and reducing it is the fastest available improvement.
- elicit before you explain. Ask what they already know first; explaining something the class could have produced wastes the best part of the lesson.
- a lesson has a shape, and students follow structure more than content.
- context before form. Show the language in a situation before naming the rule, or the rule attaches to nothing.
- correct at the end, not during fluency. Interrupting a speaking activity to correct stops the activity and teaches avoidance.
- plan the practice, not the presentation. What students do is what they learn, and the presentation is the shortest part of a good lesson.
Two markets, and only one of them has a ceiling
On a marketplace you are one row in a grid, next to several thousand other rows, and the grid is sorted by price and rating. Supply is global, so the rate converges downward towards whatever the cheapest acceptable teacher will accept, and the filters buyers use include nationality and accent, which is a fact about buyers rather than about teaching. The platform owns how you were found, so a student who stops booking has not left you, they have left a search result.
Teaching your own students inverts every one of those. The rate is set by one person's local economy rather than by a global average, the comparison is not a grid but whether the last lesson helped, and a student who is satisfied brings you the next two at no cost. The work that is different is discovery, which the platform was doing for its commission, and it turns out to be the whole job rather than a small one. That is what platforms, rates, and finding your own students is really about, and the second half of that title is the half nobody prepares for.
Fully booked at a rate that cannot move
This is the stall, and it is disguised as success. Twenty-five teaching hours a week, a full calendar, good reviews, and an income that has not changed in a year. Every remaining lever is more hours, and there are not more hours, because preparation and messages already fill the gaps between the lessons.
The mechanism is worth seeing clearly, because it is not anybody's fault. The commission on most marketplaces falls as you accumulate hours, so the longer you stay the more expensive it becomes to leave, and the moment you finally have the reputation to go independent is the moment your effective rate on the platform is at its best. Meanwhile the activity that would change anything, finding your own students, needs unbooked hours, and being fully booked is precisely the state that removes them. The way out is deliberate and it hurts: keep three hours a week unsold, permanently, and spend them on the thing that has no immediate income attached to it.
The order, with no dates
How a teaching income actually gets built
There are deliberately no timescales here. The honest range is one year to five and almost all of the variance sits at the third stop.
- startA certificate that clears filtersAsked for at the beginning, and only by employers01
- 02The first fifty paid hoursCheap and fast, and the hours matter more than the money
- A niche narrow enough to describeNurses preparing for an exam, not general English03
- 04Students who arrive recommendedThe first ones you did not pay a commission to meet
- A booking page of your ownSomewhere for a recommendation to land05
- 06A waiting list, and a rate that movesThe only lever left that is not more hours
Most teachers stop at the second stop and describe it as being busy. The third is the one that changes the rate, and it is the only stop on this road that nobody else can be paid to do for you.
What to pay for, and what to set up once
Preply is the marketplace most new teachers start on in 2026, and starting there is a reasonable decision rather than a compromise: the first fifty paid hours are worth more than any qualification and it is the fastest route to them. It replaced the local language school, which took a similar share and also decided your timetable and your syllabus. Free to join, with a commission on every lesson that falls as your hours accumulate, and knowing that the fall is what makes leaving expensive later is worth more than the exact figure.
The thing to set up on the day you take your first private student is a booking page of your own. Calendly connects to your calendar, shows only the hours you actually want to teach, and takes a payment before the slot is held. It replaced agreeing times over messages and then chasing the money separately, which costs about twenty minutes per student per week and is the reason most teachers quietly stop taking private students. The free tier is enough for one calendar and one lesson length, which is all this needs for the first year.
The mistake: preparing more than you teach
A new teacher spends two hours preparing a one-hour lesson, which turns a decent rate into a poor one and nobody notices because the preparation does not appear on any invoice. Worse, most of it is the wrong work: slides, a warmer, a worksheet, materials built for a room rather than for one person who told you last week exactly what they were stuck on.
What replaces it is one lesson shape you reuse forever and ten minutes of preparation that is specific to the student. The shape is not a template to be filled in but a sequence that survives contact with a beginner and with an advanced pair, and building it once is the entire return on a shape that works with one student or with twelve. Once it exists, preparation stops being creation and becomes selection, and the effective rate roughly doubles without a single price rise.
Where to start among the four
If nothing is in place yet, the certificate is genuinely first, and the only question is which one, since the market treats them very differently and the difference is not quality: which certificate is asked for, and by whom is a question about employers rather than about training. If the first students are already booked and the lessons feel improvised, go to the shape before anything else.
The hardest lesson to teach is the first one with somebody who has no English at all, because every technique you own assumes a shared language to explain with, and the first lessons, with no language in common is where that assumption gets removed. It is also the best-paid corner of the beginner market for exactly that reason.
What none of this settles is whether teaching is the work you want or the work that was available, and those get confused most often by people who are good at it. The wider version of that question, including what a teaching year is worth on a application for something else, is choosing work, getting it, and moving up. If the students are children the craft changes completely and almost nothing above transfers, which is teaching a child without turning it into homework. And placing a new student honestly in the first twenty minutes, before selling them anything, is where you are now, and what the next step actually contains, read from the other side of the table.