An IT project combines a constraint few other industries face: the need often evolves along the way, but the underlying infrastructure must stay stable. IT project management juggles both.
Software development popularized agile precisely because traditional IT projects suffered from a recurring problem: the need expressed at project kickoff is almost never the one that actually holds value by the end. Unlike a construction site where the plan stops changing once it's set, an IT project keeps adjusting as users discover the product.
But IT isn't only product development: infrastructure migrations, ERP rollouts, and cybersecurity projects follow constraints much closer to waterfall, with strict maintenance windows and near-zero tolerance for error. The PMI even publishes a dedicated agile practice guide separate from its classic PMBOK framework, precisely to cover this range of IT contexts.
Short sprints, prioritized backlog, frequent releases to validate with real users instead of on paper.
A cutover plan rehearsed and validated before deployment night, with a documented rollback plan.
A continuous flow of tickets, no sprint concept, suited to an unpredictable volume of requests.
Milestones and governance at the program level, agile execution inside each team.
Common mistake: defaulting to agile for an infrastructure migration project. A datacenter cutover doesn't need sprints — it needs a detailed cutover plan and a tested rollback.
Should be visible in the backlog just like a feature, not handled on the side.
A late third-party API can block several teams — track it via the critical path, not at the last minute.
Tests passing, code review, security checked — see the Definition of Done.
Backend, frontend, and DevOps have distinct workloads to track separately.
See also the Scrum Guide for the agile product-development side.
Kanban, sprints, Gantt and RACI in the same tool — switch method without switching tools.