I started on the web a long time ago, back when an online store selling supplements was still a novelty and half the internet was pushing green tea extract. I was in that first wave, building for small businesses and then for much larger ones, and I have been building on the web ever since. Long enough to have watched the whole stack turn over three or four times, and to have formed firm opinions about which parts of it were ever worth keeping.
Eventually I was running delivery for a full web team, front-end developers, back-end developers and designers, coordinating concurrent builds with the client in the room while we did it. Content management systems, customer databases, storefronts here and overseas. That is where I learned the thing everything since has been built on: the hard part of a project is almost never the code.
So what is it? It is the people. A build has trades, vendors, a client’s own staff and somebody’s boss in the room, and not one of them reports to you. It is the date, which is a promise other people have arranged their own work around, so moving it costs more than the work it was holding. It is the money, and a budget survives the change orders you actually assess and dies on the ones waved through. And it is the telling. Bad news early is a problem to solve. Bad news late is a crisis to survive, and most of the job is making sure nobody hears it last.
The code is the part that comes with documentation.
Then the projects got heavier
I moved into infrastructure, and the stakes changed. Eight years of capital work, running the technology and facilities scope inside buildings that could not stop working for a single day. Clinical unit builds. Surgical suite expansions. Disaster-recovery centres. Whole departments moved out, renovated and moved back while they carried on treating people.
More than a hundred projects. A cumulative portfolio north of fifty million dollars. Usually ten or more of them live at once, each with its own drawings, its own trades and its own date that would not move.
The work is in the detail. Reviewing issued-for-construction drawings so the data, power and low-voltage are right before anybody closes a wall. Tendering and bid evaluation on scopes up to twenty million under public procurement rules. Walking a site to find the deficiencies myself, a photo and a location for every one, rather than taking a contractor’s word that they had been fixed. Evaluating a change order against the original drawings and the contractor’s own tender rates, because a change order is a claim to assess, not a bill to approve.