There is a conversation that happens in staff rooms every September. Two teachers compare timetables. One has five consecutive periods on Tuesday and every Friday afternoon; the other has a spread with gaps in sensible places and Friday afternoon free. Both timetables are entirely valid. No clashes, no double-bookings, everything adds up. And one of those teachers is going to spend the year quietly resentful.
This is the failure mode nobody plans for, because "valid" is where most timetable builds stop. The grid fills, the clashes clear, and the process ends. Whether the result is fair was never a question anyone asked — it is simply what fell out of the order in which lessons happened to be placed.
Fairness in a timetable is not one thing. It has at least four dimensions, and they trade against each other.
The first is total contact load: how many periods each teacher actually teaches. This is the one schools do track, usually against contracted hours, and it is the easiest to get roughly right. It is also the least informative on its own, because two teachers with identical period counts can have wildly different weeks.
The second is distribution: whether those periods are spread across the week or bunched. Five consecutive periods is not the same job as five periods spread across a day with breaks between them, even though the number is identical. Teaching is a performance, and performing for five hours without a pause degrades by the fourth hour in a way that shows up in the classroom.
The third is gap quality. A teacher with three free periods scattered as single gaps between lessons has, in practice, almost no usable time — a single period is enough to answer email and not enough to plan, mark, or meet a parent. The same three periods as one contiguous block is a genuinely useful afternoon. Two timetables can show identical free-period counts and offer completely different working lives.
The fourth is the unglamorous rota: early starts, last periods on Friday, the difficult groups, the far-side-of-campus rooms. Someone has to take these. Fairness is not that nobody does — it is that the same person does not do all of them, every week, all year, because the grid happened to work out that way.
The reason manual timetabling handles this badly is not carelessness. It is that fairness is a global property, and manual building is a local process. When you place a lesson, you are solving the immediate problem: this class needs Maths, this teacher is free, this room is empty. You are not simultaneously evaluating what that placement does to the fifth-period distribution of forty staff across a week. No human can hold that. So the fairness of the final timetable is essentially an accident.
This is exactly the kind of problem automation is suited to, and it is worth being precise about why. Finding a valid timetable is the easy part; for most schools thousands of valid timetables exist. The useful work is choosing a good one from among them — evaluating a large number of valid arrangements against workload criteria and selecting the most balanced. That is tedious, mechanical, and something a computer does without getting bored. It is the strongest argument for automated scheduling, and it is not the one usually made.
Some practical advice regardless of the tooling. Decide what fairness means at your school before you build, and write it down — a maximum on consecutive periods, a minimum on contiguous non-contact time, a rota for last-period-Friday. Undefined fairness cannot be optimised for and cannot be defended when challenged. Then publish the criteria alongside the timetable. Most staff resentment is not about having a hard week; it is about believing the hard week was arbitrary. A stated rule that produced an uncomfortable result is a very different conversation from an unexplained grid.
And leave room for judgement. A teacher returning from illness, one carrying a heavy pastoral load, one in their first year — these are contextual factors no scheduling engine can see. The right shape is that software produces a balanced baseline and a human adjusts it for circumstances, rather than a human building from scratch and hoping the balance emerges.
Skoolia’s scheduling optimises for these dimensions rather than stopping at valid: it balances contact load across qualified staff, spreads subjects rather than stacking them, and works to minimise fragmented gaps. You can lock the placements you have decided on for human reasons and let it rearrange around them.
The clashes in a timetable get discovered in week one and fixed by Friday. The unfairness lasts the whole year, and nobody files a ticket about it.