Three builders come out to quote for the extension.
Two of them land within a few thousand of each other. The third comes in thirty percent under both.
You already know which one gets the job.
Here is the part worth looking at. The cheap builder is not incompetent and he is not guessing. He has built forty of these. He knows the ground behind the houses on that street, he knows what happens when you take a wall out of a building of that age, and he knows there is a fair chance of needing to underpin it.
He also knows that if he prices the underpinning, he does not get the work.
So he quotes for the version of the job where none of that happens.
Nothing in the quote is false. Every line of it is accurate, for the extension where the ground behaves, and the wall holds, and the weather cooperates. It is simply not a forecast. It is the lowest number that is still defensible on paper.
The Same Thing, with a Statement of Work
Now put that in a tender.
A software vendor is bidding for a piece of delivery work. The people writing the bid have done this before, many times, and they know exactly where it goes wrong. Interfaces to a system nobody has documented since 2011. A data migration that will turn out to be worse than the client believes. A client team that will not be available in the volume the plan assumes, because it never is.
All of it is known. Some of it is even written down internally.
None of it goes in the price.
Not because anybody is confused about whether these things happen. They are priced out deliberately, because the bid is a competition, and in that competition the honest number loses to the optimistic one every single time. Pricing risk is unilateral disarmament.
So the bid goes in on the best case, and it wins, and everybody shakes hands.
It is not technically a lie. It is purposely ignoring reality, which is a different offense and, in some ways, a worse one, because it requires knowing better.
What that Buys, on Both Sides
Then the risks land, as some of them always will.
The vendor discovers the project is no longer profitable, somewhere around month five, and the tone of the correspondence changes. Every request from the client becomes a change request. Every change request becomes a negotiation. The account manager starts attending meetings he never used to attend.
The client, who has a fixed price in writing, sees a supplier trying to charge extra for work they thought they had bought.
Both are behaving rationally. Both are furious. And the argument that follows is never really about the interface or the migration. It is about a number that was agreed a year earlier by people who knew it was optimistic and signed it anyway.
The negotiation ends in one of two places. Less scope for the same money, or the same scope for more. Neither is what anybody bought, and everybody ducks for cover.
Shift Focus from Quote to Risks
The fix is not a procurement rule. It is a question, and it belongs in the tender, where it changes who wins.
Ask every bidder for their top five delivery risks, with a price against each one and a plain statement of who carries it. Not a page of general assumptions and exclusions. Five named risks, five numbers, five owners.
The bidder who cannot answer has not thought about it and will find out at your expense.
The bidder who answers with disclaimers has thought about it carefully and is hoping you will not read them.
The bidder who shows you five priced risks has just handed you the most useful document in the entire bid pack, and will probably look expensive next to the others.
Then compare the bids on their assumptions rather than on their totals. That is the only comparison that means anything, because that is where the bids actually differ.
The gap between the cheap quote and the other two is not a discount.
It is a list of things somebody has decided not to mention yet.
The wall needed underpinning.
It was always going to need underpinning.
The quote was accurate to the penny, right up until the morning somebody started digging.