One-on-one tutoring gives the highest personalization but caps out around 8 to 12 students per tutor per week. Group cohorts multiply revenue per tutor-hour, but only work well when students are level-matched within a cohort. Most growing SAT institutes end up running both, not choosing one model exclusively.
This guide covers what each model actually earns per tutor hour, when 1-on-1 genuinely makes more sense, and how most institutes structure the mix in practice.
The math behind group tutoring's appeal is straightforward once the numbers are laid out. A tutor charging $80 per hour for 1-on-1 sessions earns $80 for that hour of work, regardless of how skilled they are or how in demand their time is. That ceiling is fixed by the format itself, not by the tutor's ability.
The same tutor running a group of 6 students, each paying 60 to 70 percent of that $80 rate, charges each student $48 to $56 per hour. Multiplied across 6 students, that same hour of teaching now generates $288 to $336 in revenue, roughly 3.5 to 4 times what the identical hour earns as a 1-on-1 session.
Where the Extra Margin Actually Comes From, and What It Costs in Personalization
That extra margin is not free. It comes directly from splitting the tutor's attention across more students in the same hour, which is exactly why group tutoring only works when students are close enough in level that the same explanation serves the whole room. A group with a wide score spread forces the tutor to either bore the strongest students or lose the weakest ones, and the margin advantage stops mattering if students start leaving because the pace does not fit them.
One-on-one sessions do not have this constraint. Every minute of the session is calibrated to exactly one student's specific gaps, which is precisely why the format still exists and still commands a premium, even though it caps revenue per hour far below what a well-run group can generate.
For Educational Institutions: An AI System to 3X Your Revenue
Generate leads and improve conversions, while reducing operational overheads - with VEGA AI
High Scorers Closing Small, Individual Gaps
Students approaching 1400 and above on the SAT are usually not missing broad content categories anymore. Their remaining gaps tend to be highly specific and highly individual, one student losing points on a narrow reading inference pattern, another on a specific Advanced Math topic that rarely comes up.
Putting these students in a group wastes the format's main advantage. There is no shared gap for the group to work on together, which means the tutor ends up running what is effectively several 1-on-1 conversations inside one session anyway, without the pricing to match. For students in this range, 1-on-1 or a very small group of 3 to 4 is usually the more honest fit.
New Institutes Without Enough Students Yet to Fill a Level-Matched Cohort
Group tutoring depends on having enough students at a similar level, taking the test around the same time, to fill a cohort of 5 to 8. A new institute with 6 total students spread across three different score ranges and three different test dates cannot actually run a well-matched group yet, no matter how appealing the per-hour economics look on paper.
In this situation, 1-on-1 is not a fallback, it is the correct model until enrollment reaches the point where level-matched cohorts become possible. Forcing a mismatched group together to chase the margin advantage usually costs more in student dissatisfaction and churn than it earns in extra revenue.
In practice, very few institutes run one model exclusively once they pass a certain size. The common pattern is a hybrid: 1-on-1 for the diagnostic session every student starts with, 1-on-1 or small groups for students at the extreme ends of the score range, and level-matched group cohorts for the bulk of students in the middle, where the format actually works best.
This is not a compromise position, it is what the underlying economics point to once an institute has enough students to actually build proper cohorts. The diagnostic and edge cases protect quality where group format genuinely does not fit. The group cohorts capture the margin advantage everywhere else.
How Does Automation Change This Calculation?
The tutor-hour math above only covers live teaching time. It does not include the administrative work behind each model, grading practice sets, tracking individual progress, answering doubts between sessions, which scales differently depending on how many students a tutor is responsible for.
In a manual workflow, group tutoring's margin advantage gets eaten into by exactly this kind of overhead, since 6 students in a group still generate 6 sets of practice to grade and 6 individual progress profiles to track, even though they share one live session hour. Automating grading and between-session doubt clearing removes most of that per-student overhead regardless of which model an institute runs, which makes group tutoring's per-hour margin advantage closer to its true economic value, since the administrative cost that used to scale with student count no longer falls entirely on the tutor.
For the operational side of actually running a group cohort once this structure is in place, from group size to session structure to mock test cadence, see how to run group SAT classes. For the full rate structures and program pricing across both formats, see the SAT tutoring pricing guide.
To see how VEGA AI supports both 1-on-1 and group tutoring models with the same platform, explore the test prep platform, check pricing options, or book a discovery call.
Save weeks of manual work—generate complete syllabus, question banks, and assessments in minutes with VEGA AI.
