You see big hourly rates, then the doubt
The first time I started comparing online English tutoring sites, the numbers looked almost too neat: $20, $30, sometimes more, sitting there like a clear side-income plan. Then the questions kicked in. Is that rate for a 50-minute lesson or a full hour? Is it only for “trial conversions” or peak-time bonuses? And how many of those lessons do people actually get each week once the algorithm stops being generous?
The doubt usually lands in the same place: advertised pay is a headline, not a paycheck. Between empty calendar gaps, slow seasons, and the time spent messaging, prepping, and waiting on no-shows, two tutors can “earn $25/hr” on the same platform and take home completely different weekly money.
Start with your non‑negotiables before comparing platforms
Once that gap between headline rate and weekly money shows up, the comparison gets easier if the order changes: lock in your non‑negotiables first, then look for platforms that can actually meet them. The main constraint usually isn’t teaching skill—it’s time. If you can only work 6–8 hours a week, a site that “can pay $30/hr” but needs you online during peak Asia evenings (or requires minimum weekly availability) can turn into a slow, frustrating calendar.
I write mine down like filters: fixed schedule windows, minimum acceptable effective rate (after fees), student type I’m willing to teach, and how much unpaid admin I can tolerate. Training time is a cost too; a 10–15 hour onboarding makes sense if bookings are steady, but it’s expensive if you’re still hunting for students two months later.
Pay looks simple until you count unpaid time
After the non‑negotiables are on paper, the next surprise is how fast “$X per hour” turns into “$X per paid lesson.” Most platforms only pay for instruction minutes, so the real question becomes: how many minutes of unpaid work sit around each paid block? A 25‑minute lesson that pays $10 looks like $24/hour on a spreadsheet, until you notice you’re spending 5 minutes writing feedback, 5 minutes confirming next steps in chat, and another 5 minutes pulling materials because the student changed goals midweek.
The easiest way I’ve found to keep this honest is to calculate an effective hourly rate from a typical week, not from the platform’s headline. Take total payouts, then divide by total time you actually spent “on the job”: teaching time plus prep, notes, messages, tech setup, and the dead space caused by scattered bookings. If you’re logging 6 paid hours but the week consumes 9.5 real hours, your $25/hour headline is suddenly closer to $15–$16/hour before taxes. That number is what your schedule will feel.
Company-assigned students versus open marketplaces
After you’ve priced in unpaid time, the next fork in the road is how students land on your calendar. With company-assigned models, you’re mostly saying yes or no to what shows up. The upside is momentum: fewer hours spent marketing yourself, fewer awkward “why should I pick you?” chats, and a faster path to a routine. The constraint is control. If the platform prioritizes new tutors, pushes low-priced trials, or suddenly shifts demand, your weekly hours can slide without anything you did changing.
Open marketplaces flip that. You can raise rates, specialize, and build repeat students, but you’re also running a tiny funnel: profile tweaks, response speed, intro messages, and sometimes discounted first lessons. Early on, bookings often cluster and then stall, which makes planning around a day job harder. If your income needs to be predictable, assignment systems usually feel steadier; if your goal is to grow into higher rates, marketplaces can pay off—after a longer, messier ramp.
High advertised rates versus reliable bookings
Once you’ve felt the difference between assigned momentum and a marketplace ramp, the next reality check is how often the calendar actually fills. A platform can advertise $30–$40 an hour and still leave you staring at open slots because demand is thin, your time zone is off-peak, or the algorithm is feeding trial lessons to newer profiles. In practice, the rate only matters when it repeats, and “repeats” is a booking problem, not a teaching problem.
When I compare platforms now, I treat bookings like a reliability score. How many paid lessons per week do typical tutors report after month one? How seasonal is the demand? Are cancellations paid, partially paid, or unpaid? A site that nets $18–$22 per teaching hour but keeps 6–10 hours booked most weeks can beat a $35 headline that swings between two lessons and none. The constraint is patience: higher-rate setups often require a longer runway before the hours stabilize.
Kids-heavy platforms change your effort and churn
After bookings start looking “reliable,” the next thing that quietly changes the workload is who the students are. On kids-heavy platforms, the lesson itself can be smoother (clear curricula, supportive parents, repeat schedules), but the effort shifts into performance: higher energy, more props/slides, tighter pacing, and less tolerance for “off days” after your regular job. If you’re only teaching 2–3 nights a week, that can be fine; if you’re trying to stack 10+ hours, it can feel like sprinting.
Churn also behaves differently. Kids age out, schedules change with school terms, and parents shop quickly if a class feels dull—even when your English teaching is solid. That means more trials, more first-week handholding, and more unpaid reset time per student, which can drag down your effective hourly rate unless the platform feeds you replacements fast.
Fees, taxes, and policies that quietly cut pay
After you’ve adjusted for churn, the next hit to take-home pay is the stuff that never shows up in the hourly headline. Platform commission or “service fees” can skim a meaningful slice on marketplaces, and some payment processors add their own percentage or a fixed withdrawal charge. Even a $2 fee every time you cash out starts to matter if you’re only pulling small weekly payouts. Exchange-rate spreads can be a quiet extra cut too, depending on how they pay and how your bank receives it.
Then taxes and policy quirks pile on. As a contractor in the U.S., a good chunk of profit can disappear into self-employment tax and quarterly estimates if you don’t plan ahead. And cancellation/no-show rules are basically pay rules: some sites pay nothing, some pay partial, and some claw back bonuses if your cancellation rate ticks up. A “strict” policy can protect earnings—or make a bad week expensive.
A practical short list of best-fit websites
Once fees, taxes, and cancellation rules are in the math, the “best” site usually means “best fit.” If you want fast onboarding and a predictable floor, Cambly is the cleanest benchmark: fixed per-minute pay and no pricing decisions, but the ceiling is limited and wait time can dilute the effective rate.
If you can tolerate a ramp and want upside, marketplaces split into two practical picks: Preply if you’re okay paying commission early while you build repeats, and italki if you prefer more profile control and long-run regulars.
