Most K-12 districts pull a teacher retention number once a year — typically the percentage of teachers who returned for the following school year. It looks like a metric. It rarely changes behavior. By the time the number is in front of a superintendent, the teachers who were going to leave have already left, and the conversation shifts to hiring. Teacher retention metrics that matter are the ones that surface risk early enough to do something about it.
Workload metrics that predict burnout
Annual retention is a lagging indicator. The leading indicators are workload signals that drift before someone decides to leave: number of preps per teacher, total contact minutes per week, after-school duty load, and grading volume in the two weeks before report cards. Inside Borderset, these signals are calculated from the schedule and gradebook rather than a separate survey, so they reflect what is actually happening — not what teachers remember when asked in May.
The three-prep red line
A common pattern: secondary teachers with three or more distinct preps in a single term churn at noticeably higher rates than peers carrying one or two. The mechanism is straightforward — prep load compounds with grading load. Surfacing prep count alongside section size lets principals rebalance assignments before April, not after. The same view exists on the teacher tracking module so HR and academics see the same numbers.
Year-2 retention as the real test
Most attrition discussions focus on year-1, but year-2 retention is often the more honest signal. Teachers who survive the first year and still leave after the second usually leave because of structural fit issues — scheduling, coaching, or workload — that the school can change. Borderset retention reports break the cohort apart by years of service so leaders can see whether the problem is recruitment or the second-year experience.
Exit-survey patterns to actually track
Exit interviews tend to collect honest answers and then sit in a folder. The fix is to code exits into a small, fixed taxonomy — workload, compensation, leadership, commute/life, classroom assignment, growth opportunity — and watch the distribution change quarter over quarter. When one category climbs across two consecutive cycles, you have a system issue, not a personal one. Tie exit codes back to the teacher record so the next round of credentials and PD tracking can prioritize the right supports.
For superintendents, the same data rolls up into the principal dashboard alongside staffing forecasts. Access is governed by role-based access so individual workload data stays visible only to people responsible for acting on it.
What to stop reporting
Drop the single-number annual retention rate from the monthly board pack. Replace it with three numbers: percentage of teachers with three or more preps this term, percentage of new teachers (years 1 and 2) on schedules above the workload median, and trailing-12-month exits by coded reason. Each of those numbers is something a principal can change in the next scheduling cycle. The single annual number is something a principal can only explain.
One more pattern worth surfacing: late-spring schedule changes. When a teacher's class assignment shifts significantly between the May draft schedule and the August final, retention risk climbs noticeably in the following year. The change itself is often defensible — a section was added, a colleague left, a program was launched — but the teacher experiences it as a quiet downgrade of their role. Tracking the count and magnitude of late schedule changes per teacher gives principals a chance to have the conversation before September, when the most predictable response (resignation) becomes a hiring problem instead of a coaching one. Borderset logs schedule revisions so the May-to-August delta is visible on the teacher record, not hidden inside a master schedule export.
Retention is a scheduling problem and a coaching problem more often than it is a pay problem. Borderset surfaces the scheduling half inside the workflow teachers already use, so the conversation can move from explaining attrition to preventing it.