Dariusz Kubies MCIOB · 3 September 2026
2 min read
Forty-four specialist contractors, from the United States, Europe and the UK. One of them was there to make lightning.
Science Museum, Wonderlab, 2016. I was the project manager.
Asbestos removal first, then strip out, then a full refurbishment to build the gallery, including a bespoke suspended auditorium. All of it inside a museum that stayed open and kept running around us.
Forty-four is the number people react to. It is not the hard number.
The hard number is how many of them need the same part of the same floor in the same week.
Because a gallery is not a Cat B office. On an office you hand over floor boxes, a ceiling grid and a set of keys, and somebody else brings the desks in afterwards. Here the building work was the thing that forty-four other companies were going to install into.
So the exhibits were never the difficult part of that job. The services underneath them were.
An installer arrives with a date, a crate and a set of requirements, and by then your floor is down, your ceiling is closed and your containment is in. Put a service in the wrong place and you are opening it all back up in front of forty-three other people who each have their own date.
And the risk was not sitting inside any one package. It was sitting between them.
Which is the part I still think most programmes get wrong. We are good at the durations inside a trade and we are careless about the fortnight where two trades need the same ceiling void. The bar looks fine. The interface is where the job actually goes wrong.
The other thing that job taught me is that you do not hand a gallery over and go home. We coordinated the specialist installers and then stayed through their testing phase, which is a different job again. You sit in a building you have finished while somebody finds out whether the lightning works.
Wonderlab is still open.
For anyone running something with a lot of specialist interfaces at the moment. Where do you actually hold that risk on your programme, and does it have a duration against it or is it just the gap between two bars?
A simple way to hold that risk on a programme:
List the shared spaces: the ceiling void, the riser, the floor zone.
For each one, name every trade that needs it and the order they need it in.
Give the handover between two trades its own task, with its own duration and an owner.
Check each handover against the date the next installer says they are arriving.
It is simple. The hard part is remembering to do it, because the gap between two bars has no name on the programme.
Related: A Programme Can Look Solid and Still Be Fundamentally Weak
Did this resonate? I read every comment.
Comment on LinkedIn →
Dariusz Kubies MCIOB
Founder, FitOut Insider · Project manager, MCIOB · 20+ years in UK fit-out
About Dariusz →