Priced by nobody, built by somebody
The structural steel package ends at the underside of the tank. The mechanical package starts at the tank shell. Between them are the brackets, the shims and the grouting that hold one to the other.
Both bills were prepared carefully. The steel bill covers every member the steel contractor is building; the mechanical bill covers every item the mechanical contractor is installing. Neither contains the connection, because from inside each package it belongs to the other one.
Nobody was careless. Two scopes can each be complete and correct and still not cover the whole, and that is a different failure from anybody making a mistake.
Why no register holds it
Risk Week 3 builds the risk breakdown structure as a decomposition of the works: take the scope, break it into parts, and identify what can go wrong in each part. The structure is sound and it is the reason the method works.
It also fixes what the register can see. Every branch of an RBS hangs off a piece of scope, so an item belonging to no scope has nowhere on the tree to sit. It isn't ranked low. It is absent.
Now run that across a multi-package job. The steel contractor's register covers the steel scope, thoroughly. The mechanical contractor's covers theirs, thoroughly. Both are complete against their own scope, both would pass an audit, and the brackets appear in neither.
Which is the uncomfortable part. This isn't a gap that better risk management inside each package would close, because the discipline that makes each register good is exactly the discipline that excludes the item.
Where the gaps sit
They are not random, and knowing the pattern is most of the work.
Connections between two systems, where each package delivers to a boundary and something has to join them.
Temporary works serving more than one party — access scaffolding, craneage, a haul road. Each package prices what it needs alone; nobody prices what two need together.
Protection of finished work while another trade operates around it. The party who owns the work has finished; the party causing the risk has no obligation towards it.
Testing and commissioning across a boundary, where each side can prove its own portion and neither can prove the combination.
The final surface where two trades meet, which is a scope question on every project ever built.
Who finds it, and when
The discovery route is the same every time and it is the worst possible one.
The item is found by the first person who has to build across the line, at the moment they have to build across it. Not in a review, not in a register, not at tender. On the day, in the field, when both parties are already on site and both are already busy.
At that point it is a change with no budget, no owner and no obvious payer. The steel contractor will price it as a variation, because it isn't in their contract. So will the mechanical contractor. Neither is being opportunistic; the work genuinely is outside what either agreed to do.
And the employer pays, because week 2 established where the integration risk went when the packages were let. This is that risk arriving in its most concrete form: work that exists physically and existed in no price.
The register that has to be invented
Since no existing register can hold these items, one has to be made whose organising principle is the boundary rather than the scope.
One line per pair of packages that touch. Not per package — per pair, because the item belongs to the relationship and not to either party. Against each line: what physically happens where they meet, which contract prices it, and if the answer is neither, who is going to.
Reporting Week 24 separated registers that track state from inventories that record what exists. This one is a state register, it feeds the constraint log the moment an entry has no answer, and it is the only place on the project where these items can live.
When to do it
The value of the review collapses with time and the curve is steep.
Before award, an unpriced item is a scope clarification: it goes into one package or the other, at tender rates, and costs the difference between two prices. After award it is a variation, at variation rates, with the leverage on the other side. On the day it is discovered in the field it is a variation and a delay.
So the review belongs at the point where the scope splits are being drawn, and it is cheap there — a day with the boundary drawings and somebody from each discipline. The same exercise after the packages are let is worth doing and worth considerably less.
System design
One register that no existing template will give you, because every template is organised by scope and this one is organised by boundary.
| Record | Produced by | Required quality | Verified against | Feeds |
|---|---|---|---|---|
| Boundary register | Project controls | One line per pair of packages that touch, not per package | The scope split drawings | Scope gap list · constraint log |
| Scope gap list | Project controls with engineering | What physically happens there and which bill prices it | Both bills, side by side | Clarification or variation |
| Pricing owner | Commercial | Named for every item where the answer was neither | The award decision | Budget · change register |
| Shared temporary works | Construction | Listed where more than one party needs the same thing | Method statements | Cost allocation · programme |
The third row is what stops the list becoming a record of problems nobody solved. An item with no pricing owner isn't a scope gap that has been identified; it is a scope gap that has been written down, which is a different and much weaker thing.
Practical insight
Take the two packages on your own project that physically touch the most, and put the two scope documents in front of you side by side.
Then walk the boundary instead of reading. Follow the line where one of your packages ends and the next begins, and every time something crosses it — a connection, a service, a support, a surface — ask which of the two bills in your hands contains it.
You will find items in both, and that is a clash you can resolve in a morning. You will find items in neither, and that is what this week is about. Write your list today, while every line on it is still a clarification. The same list in six months is your schedule of variations, and by then you are on the paying side of it.
Key takeaways
- Two scopes can each be complete and correct and still not cover the whole.
- A risk breakdown structure decomposes a scope, so an item in no scope has nowhere on the tree to sit.
- The item is absent rather than ranked low, and better risk management inside each package won't find it.
- The gaps are patterned: connections, shared temporary works, protection, cross-boundary testing, and the final surface.
- They are found by the first person who has to build across the line, on the day they have to.
- By then it is a variation from both sides, and neither party is being opportunistic.
- The employer pays, because that is where the integration risk went when the packages were let.
- A boundary register is organised per pair of packages, not per package. The item belongs to the relationship.
Records born here. The boundary register, one line per pair of packages · the scope gap list with who will price each item · the clarifications issued before award.
What is coming next
A register of boundaries is a document. Somebody has to own the boundaries themselves, and that is a job with a name on some projects and nobody's name on most.
Next week: interface management as a function — owning a boundary rather than reporting one.
Enjoyed this lesson?
Join with Google to get each new lesson the moment it's published — and help me see which topics matter most to you. No spam, one email a week, unsubscribe anytime.
Already following on LinkedIn works too — this is just for the weekly email.