The expensive mistake in tendering is not losing. It is losing slowly — spending three weeks of estimating capacity on a tender the firm was never positioned to win, and then submitting anyway because of the effort already sunk into it.
A bid/no-bid decision made on day one costs twenty minutes. The same decision made on day nineteen costs the three weeks, plus the opportunity cost of the tender you did not chase.
Six questions, in order
Order matters here. The first three are eliminating questions — a bad answer ends the review, and you stop reading. The last three are scoring questions, which only matter once the eliminating ones are clear.
1. Are we eligible at all? Licensing, qualification, bonding capacity, prequalification status, mandatory site visit already past. This is binary, it is knowable in minutes, and it is the most common reason a firm should have stopped and didn't.
2. Can we physically make the deadline? Not "is it tight" — can the estimating team produce a defensible number by the close, given everything else already committed. A date on a calendar is not capacity.
3. Is the work actually our work? Self-perform scope, geography, project size relative to the firm's normal range. A project three times larger than anything the firm has delivered is a different business, not a bigger version of the same one.
4. Who else is bidding, and who holds the incumbency? An incumbent with a clean performance record on a straight low-bid tender is a strong signal. Not decisive, but it should change the expected value of the effort.
5. How much risk has the owner transferred? Liquidated damages, unusually short payment terms, unpriceable allowances, one-sided change-order provisions. These are visible in the front end of the tender well before the drawings are understood.
6. What does it cost us to bid? Estimating hours, bond cost, any specialist input. Compare against a realistic win probability, not an optimistic one.
Why it does not happen
Every firm agrees with this and most do not do it. The reason is not discipline — it is that answering the first three questions properly means reading the instructions to bidders, the supplementary conditions, and the special provisions, spread across several hundred pages, before you have decided the tender is worth reading at all.
So the review gets deferred. The estimator starts on the takeoff because the takeoff is the part they can start, and the triage happens implicitly, three weeks later, by which point the decision has already been made by accumulated effort.
Front-load the reading, not the pricing
The practical fix is to separate the two kinds of reading. The eliminating questions live in a predictable part of the tender — the front end, not the drawings — and they can be answered before a single quantity is taken off.
This is one of the things automated extraction is genuinely good at. Pulling every deadline, mandatory requirement, licensing condition and risk-transfer clause out of a tender set is mechanical work, and having it in front of you on day one changes the twenty-minute conversation from a guess into a decision.
The goal is not to bid less. It is to spend the same estimating hours on tenders the firm can actually win — and to make that call while it is still cheap to make.
Related reading: the five documents that sink a public tender.
