Four sessions,
and then a real conversation.
Most tutor onboarding consists of handing over a name, a time and a subject. It is not enough, and the gap shows up in the first session, which is the one that decides whether the family books a second block. Before a new tutor meets a student they need three things: what the student is actually trying to achieve and by when, what has already been tried and did not work, and anything about the person — anxious about the subject, recently changed school, does not like being asked questions in front of a parent. Five minutes of briefing prevents a first session spent discovering what you already knew.
Then structure the first four. Session one is diagnostic: find out where the student really is before teaching anything, which is the habit most new tutors skip because they want to be useful immediately. Sessions two and three are the work. After session four, sit down with the tutor properly — not a passing "how's it going", which produces "fine" — and ask three things: what is this student's actual problem, what are you doing about it, and what do you need from me. The answers tell you almost everything about whether this hire is working.
Tell the family the same rhythm. A short note after the first session saying what the tutor found, and a fuller one after the fourth, converts an anxious start into a visible process — and it is the moment when most parents decide whether this centre is competent.
Standardise the promise.
Not the teaching.
The principle underneath the table: standardise what the family experiences, leave alone what happens inside the teaching. A parent should get the same shape of communication, the same first-session structure and the same cancellation policy whichever tutor they were assigned, because those constitute the promise the centre is making. How a tutor explains simultaneous equations is not the promise; it is the craft you hired, and prescribing it produces worse teaching from good people.
Over-standardising is the more common failure once a centre starts growing, and it is usually a response to one bad experience. Somebody taught badly, so now there is a mandatory lesson template, and within a term the best tutors — who had their own better approach — have left or quietly ignore it. Fix the specific problem with the specific person instead.
Three cheap habits
that beat any framework.
Read what students produced. Not the tutor's account of the session — the actual work. A tutor reporting steady progress while the student's written work has not changed in six weeks is the commonest quality problem in tutoring, and it is invisible unless somebody looks at the work.
Join part of a session, as routine. Fifteen minutes, occasionally, for everyone including your best people, framed from day one as normal practice rather than as a response to concern. If observation only ever happens when something is wrong, it becomes an accusation and tutors perform rather than teach. Online this is easier than in a room, and it is worth using that.
Read the notes going to parents. They are the most reliable early indicator you have. Notes that are vague, identical week to week, or late are almost always a sign of a tutor who is not really engaging with that student — and because the parent is reading the same thing, you are seeing what they see.
Rebooking rate,
read carefully.
One number cannot be argued with: the proportion of students who book another block after finishing one, per tutor. Families vote with money, and they do it quietly, so this is where the silence becomes legible. If one tutor is at a materially lower rate than the others over a reasonable number of students, something is happening that nobody has told you about.
Read it fairly, though, because the naive version punishes exactly the wrong people. A tutor given the most difficult students, the ones already close to giving up, will show a lower rate than one handed motivated exam candidates — and if you act on the raw number you will push your strongest tutor towards easier work or out of the door. Look at who they were given before you look at what happened. Also allow for volume: three students is not a rate, it is an anecdote.
This is one thing a platform genuinely helps with, not by measuring quality but by making the underlying facts retrievable. If sessions, attendance and packages sit in one record rather than a diary and a bank statement, then which students a tutor taught and whether those students came back is something you can look at in a minute rather than reconstruct from memory once a year. The packages article covers the block structure that makes a rebooking rate a real number in the first place.
A skill gap and a fit problem
need different answers.
When quality is not where it should be, work out which of two things you have. A skill gap is a tutor who wants to do well and does not yet know how — they explain one way, or they teach without diagnosing, or their notes are thin. That is coachable, usually quickly, and the conversation should be specific and practical: here is what I saw, here is what to do instead, let us look again in three sessions. Vague feedback produces no change and damages the relationship.
A fit problem is different: someone who does not enjoy the part of the job that matters, or does not much like the students, or is unreliable in a way that does not respond to being asked. That is not coachable and is unkind to prolong, for them and for the families. Small centres routinely carry this for a year because the alternative conversation is unpleasant and because the tutor is a nice person — while families quietly leave and take the reputation with them. Decide early, say it clearly, and be decent about it.