I've delivered platforms to government ministries and walked away. I've trained teams, handed over documentation, transferred repositories, and moved on to the next project. What I've noticed, years later, is that some of those systems are still running. Others are not.
The difference is almost never technical quality.
What actually persists
They were built around people, not processes. The best govtech I've seen isn't the most sophisticated — it's the most adopted. Adoption comes from understanding the actual humans who will use the system. A platform that staff feel ownership over survives budget cycles, leadership changes, and vendor transitions. One that was imposed on them doesn't.
The institution knows what it has. Documentation isn't optional. It is the mechanism by which a platform survives the departure of the people who built it. When I leave a project, I want the agency's own IT team to be able to answer "why does it work this way?" without calling me. If they can't, I haven't finished the job.
The architecture has room to breathe. Systems built to be unmaintainable — tightly coupled, undocumented, vendor-dependent — are designed to fail on a timer. The ones that last are modular enough that someone who wasn't there at the beginning can understand, extend, and fix them.
" Indispensability is the enemy of longevity. The best thing you can build is something that works brilliantly without you.
The harder question
Why does it matter whether the thing outlasts you? I think building things that outlast you is an act of a particular kind of faith — the belief that your work is genuinely useful, not just to a present need but to a future you won't see. That orientation changes how you make decisions. You document more carefully. You transfer knowledge more deliberately. You resist the temptation to make yourself indispensable.
What I've learned to look for
When I assess a potential project now, I try to understand early whether the client actually wants something that will outlast the engagement — or just something that will check a box in a quarterly review. The honest answer is that many clients, knowingly or not, want the latter.
I take fewer projects now. The ones I take, I try to build to outlast me. Not because it earns more — it often earns less. But because building things designed to be temporary feels like a waste of the only currency I'm actually spending: time.
" Build things worth outlasting you. The rest follows.