Why modernize
The biggest decision your school makes twice.
Once when you build it, and again every time a student's request quietly goes unmet. Here is what the old way actually costs — with sources, not slogans.
A schedule is a budget you spend in people
Every section you open is a teacher's time. Every singleton you place rules out every other singleton in that period. Every unfilled request is a student who takes something else instead.
What building it by hand costs
The most careful public study of this is Columbia Law School's Center for Public Research & Leadership report About Time: Master Scheduling and Equity. Its findings are unusually concrete.
100+
Hours of "pretty much nonstop work" for a single scheduler, spread over multiple weeks.
Columbia CPRL, About Time, p.9 (2021)
1
Person. "The bulk of master scheduling is too often left to a lone programmer who independently works out the complex puzzle." The report recommends a team of three to five.
Columbia CPRL, About Time (2021)
14 mo
A realistic build timeline runs planning in August–December, course selection January–March, building April–May, and fine-tuning June–October.
ACSA, Building a Master Schedule (2018)
The work is enormous, it lands on one person, and it stretches across most of a school year — so the schedule is still being revised long after it should have been finished.
"There was about two to four weeks at the start of the school year… where students and teachers didn't have schedules. Some kids were in the gym, some kids were in the cafeteria, and some kids weren't in classes for a whole month."
Vice Principal Diane Conti, Herbert Hoover High School, San Diego USD — quoted in Columbia CPRL, About Time, p.49
The old tools check the wrong things
The same Columbia report surveyed eight scheduling tools and SIS scheduling modules to see what each one could actually catch. Conflict detection was near-universal. Almost nothing else was.
| Does the tool alert you when… | Of 8 tools surveyed |
|---|---|
| Two courses conflict | 7 of 8 |
| A class is demographically imbalanced | 4 of 8 |
| A student is scheduled into a course they already took | 4 of 8 |
| A student is missing a course needed for graduation | 4 of 8 |
Source: Columbia CPRL, About Time: Master Scheduling and Equity, Appendix A Landscape Analysis, p.45 (2021). The report does not name which eight tools were surveyed.
"Without effective tools, schools relied on human reviewers of the schedule to find and report errors. Sometimes this meant that students did not receive their appropriate set of classes until months into the school year."
Columbia CPRL, About Time, p.17
Scheduling is an equity decision, whether or not you treat it as one
The reason this matters beyond operational tidiness is that schedule errors are not randomly distributed. Students who need the most support tend to be the ones with the most constrained schedules — pull-outs, intervention blocks, support services — which means they are the hardest to place and the first to get whatever is left.
What nOS does about this
nOS reports fulfillment broken down by grade, lists the top-missed courses, and attaches a specific reason to every unmet request — not a status code, an explanation of what stood in the way. You can see which groups of students are absorbing the compromises before you publish, not in October.
What actually changes
Building it by hand
- One pass. Trying a different approach means rebuilding from the start, so you rarely do.
- You find out a placement was wrong when a person notices.
- The reasoning lives in one person's head and a binder.
- "What if we moved AP Bio to third period" is a week of work, so it stays hypothetical.
- Undo is a photograph of the board you hopefully took.
With a solver underneath
- Thousands of requests placed in minutes, so a second and third attempt cost nothing.
- Conflicts surface as you move a section, before you let go of it.
- Every unplaced request carries its own reason.
- Scenarios run side by side and you promote the one that wins.
- Changes are staged, previewed with their impact, and committed only when you say so.
Austin ISD moved 33 schools off magnetic whiteboards and reported saving at least 50 hours and about $2,600 per campus — and described the shift in terms that have nothing to do with hours: staff "felt like they could actually manage master scheduling, and the master scheduling wasn't managing them."
"But we manage"
We've always done it this way and it works
It probably does work, in the sense that school opens and students have schedules. The question worth asking is what it costs to make it work — how many evenings, how many requests you stopped trying to fill because there was no time left, and how much of the method exists only in one person's head. Those costs are real; they are just not on any line item.
Our school is too small to need software
Small schools have the harder scheduling problem, not the easier one. Fewer sections means less room to move things, and more courses that run exactly once — and a singleton that collides with another singleton has no fix except a redesign. The case for small schools is its own page.
We already have a scheduler in our SIS
Then keep it for what it is good at. Your SIS is the system of record for enrollment, grades, and reporting, and nothing here suggests replacing it. The scheduling module inside it is a different question. Keep the SIS, replace the scheduler →
Switching mid-cycle sounds impossible
It is the most common real objection, and the honest answer is that timing matters. Here is what the move actually takes, week by week.