Single Blog

The efficiency conversation in this industry is almost always about site: better scheduling software, better plant tracking, better reporting. Our own numbers say the money is somewhere else entirely.

Where the waste is

We went back through eighteen months of variation orders on four schemes and sorted them by root cause. Just over sixty percent traced to a coordination failure that existed in the drawings before anyone broke ground — a duct through a beam, a riser that did not line up floor to floor, a door swing into a structural column.

Each of those cost between two and nine days on site plus the abortive work. None of them cost anything to fix in week four of preconstruction.

What a shared model actually buys

Clash detection is the obvious answer and it is only half of it. The larger gain is that when structure, services and architecture are in one model, the questions get asked earlier by people who would otherwise not have met.

On Horizon Station we ran weekly coordination sessions for nine weeks before the first pour. They produced 412 resolved clashes and, more usefully, three design changes that came from a services engineer noticing something in the structural sequence. Variations on that job ran at 1.4% of contract value against our average of 4.1%.

The part nobody puts in the case study

It only works if the model is the record. The moment somebody issues a revised drawing that is not in the model, the model becomes a visualisation and you are back to coordinating by email.

That is a contractual question more than a technical one. Put model status in the information release schedule, name who owns each federated file, and make the coordination session a contractual meeting rather than a courtesy.

Where to start

If you do one thing: bring the temporary works designer into preconstruction. They are the trade most often asked to solve on site a problem that was created on a drawing, and they will find it in an afternoon.