For ten years, I served as Head of Technology at an independent school — a role I never set out to grow, but one that grew anyway, one need at a time. I inherited a single administrative system running on a mainframe with dumb terminals, kept alive by Unix scripts I wrote myself, serving a handful of administrators and one computer lab. Over the next decade, that grew into something else entirely: more labs, more classrooms wanting access, and eventually connectivity itself becoming something the school needed to take seriously. Somewhere in that growth, it became clear we'd been making decisions one at a time without any real direction behind them.

Three things were converging at once. Financially, we'd never really priced technology honestly — not just the purchase, but the maintenance, the support, the wiring a spread-out campus didn't have, even the classroom space a bank of new computers would quietly take from something else. On the support side, every new system meant more complexity landing on the same support staff — me — and hiring more people wasn't a sustainable answer to a problem that kept growing. And culturally, I'd watched what happened at another school that gave every teacher a laptop with the best of intentions: instruction quality dropped, by the teachers' own account, because nobody had thought through how the technology would actually get used. Good intentions weren't a strategy. The school's head and I agreed: we needed one.

Writing it surfaced tensions that couldn't be resolved with a policy document alone. Should the school standardize on one platform — cheaper to support, easier to lock down — or let people choose the tools that fit how they actually worked, even if that meant supporting Mac and PC, Word and Google Docs, side by side? What did "support" even mean for a school — a corporate-style help desk, or something that better matched how the school actually operated? And how do you evaluate whether new technology succeeded without either rubber-stamping good intentions, the way the laptop giveaway had, or strangling the handful of genuinely innovative teachers — some of them speaking at national ed-tech conferences — who wanted room to experiment?

The plan came down to a single line we returned to constantly: all technology must support the school's culture and mission. Everything else followed from that test. We leaned toward giving people freedom over their own tools — on the condition that they owned the responsibility that came with it, learning their systems and troubleshooting their own problems. For people who found that intimidating, we built a peer mentor network — people to turn to for guidance, not a ticket queue. Every new technology request had to show how it served the mission, with a real plan for training, prototyping, and evaluation. And evaluation itself changed — from "prove you did what you said" to an honest, non-punitive reflection built around one question: what can you do to turn this into a success? On the financial side, we made the real cost of every request visible — what the school covered, and what the requester was responsible for themselves.