Open book estimate: when it works, and when it does not

An open book estimate is often sold internally as the answer to a pricing problem. The contractor opens the cost build-up, the owner gets to see the rates, and the argument about margin is supposed to end there. What usually happens instead is that the owner receives several hundred lines of cost detail and has no independent basis for deciding whether any of it is reasonable.

What open book actually gives you

Open book is not a contract type. It is a cost transparency arrangement that sits inside one, normally a reimbursable or target cost contract. In NEC4 Option C, for example, the target is the contractor's priced activity schedule, interim payment is assessed on Defined Cost plus a fee percentage, the contract names which costs are Disallowed, and the difference between the final cost and the target is split at completion through the contractor's share.

That structure gives the owner the right to see how a number was built. It does not give the owner a view on whether the number is right. Those are two different capabilities, and the second one is the one most owner teams are missing when they sign up for open book.

An open book estimate without a should-cost model is an audit of arithmetic

If the first time you form an independent view of what the work should cost is when the contractor's build-up lands on your desk, you are checking their sums against their own assumptions. The UK government guidance on should cost modelling puts the model before the market: it is to be agreed as part of the business case and procurement planning, before the contract is advertised. The same guidance uses it as a filter afterwards, with bids more than 10 percent below the model referred for review rather than accepted as good news.

The sequence matters more than the sophistication of the model. A rough independent view built before tender beats a detailed one built after the bids arrive, because the second one will be anchored on the numbers it was supposed to test.

The estimate is usually weaker than the argument about it

Opening the book does not change how well the project was defined, and definition is what sets the accuracy. Under the AACE International estimate classification for process industries, a Class 3 estimate corresponds to 10 to 40 percent project definition and carries an expected range of minus 10 to minus 20 percent on the low side and plus 10 to plus 30 percent on the high side. A Class 5 estimate, at 0 to 2 percent definition, ranges from minus 20 to minus 50 percent and plus 30 to plus 100 percent.

Read those ranges next to the typical open book review, where two teams spend a fortnight on labour rates and a productivity factor worth a few percent. That argument sits inside the noise of the estimate class. The more useful question in the room is which class this estimate is, what definition it was built on, and what has changed in the scope since it was priced. Estimate class follows where the project sits in its lifecycle, the sequence set out on our one page summary, so the honest answer is sometimes that the estimate cannot yet carry the decision being asked of it.

When an open book estimate stops working

The arrangement stops paying for itself under conditions that are visible before signature:

• Disallowed Cost is left loosely defined, so every rejection becomes a negotiation
• the owner has no cost engineering capacity to read the build-up, so open book becomes a filing exercise
• the target is set before the scope is defined, which turns the share mechanism into a bet rather than an incentive
• the scope was stable enough to competitively tender as lump sum, and open book was chosen to avoid a difficult conversation about it
• nobody is named as the person who reviews the monthly cost submission, so it is reviewed by whoever has time

None of these are contract drafting problems. They are capability and sequencing problems, which is why they show up again on the next project with a different contract form. The work sits in procurement and CAPEX strategy rather than in the legal review.

Before the next tender

Three things, in this order, and they are cheaper than any of the arguments they prevent. Build the independent cost view before the contract goes out, not after. Write the Disallowed Cost definition with the cost engineer who will actually apply it, not only with legal. Then put a name against the monthly cost review and give that person the authority to reject a line. If none of the three is in place, open book will produce transparency about a number nobody can challenge, which is worse than a lump sum, because it feels like control.