Estimating
Estimating With Evidence, Not Guesswork: Where Bid Risk Actually Hides
Three bidders priced the same public package in March 2026 and came in $1.78 million apart. Quantities rarely explain a gap like that. Here is where the risk usually sits, and how to price it with evidence instead of a contingency line.
On March 19, 2026, DFW Airport opened bids for its Data Center 1 HVAC/BAS & Electrical Rehabilitation (PA1971). Three contractors priced the same documents. The bids were $6,476,000.00, $7,381,052.28 and $8,251,823.00. The high bid was $1,775,823 above the low, about 27%.
From the outside you cannot know why. But anyone who has run a bid desk knows the usual suspects. Different reads of scope. Different views of the schedule. And different decisions about the risk in the contract terms, often made under time pressure by an experienced estimator who knew something was there but did not have time to prove it.
That last one is where guesswork lives.
Where the risk hides
Quantities come from drawings, and good estimators are very good at them. The risk that turns into guesswork usually sits somewhere else:
- In the contract terms. In this package, liquidated damages of $2,500.00 per day (Special Provisions §1.0 B), a claim notice due within 10 calendar days and before the work starts, waived if missed (General Provisions §50-17), and 5% retainage held to final payment (§100-5).
- In the specification front end. Division 01 sets how you get paid, how fast you must respond, and what you must submit. Here, an allowance price quote is due within 7 calendar days of a request (Spec 01 30 00), and the baseline schedule within 14 days of notice to proceed (Spec 01 32 16).
- In the addenda. Requirements move. In this package, Addendum 1 changed who had to hold SBE+ certification.
- In conflicts between documents, where two documents describe the same thing differently, and you have to pick one.
Take the first one. The liquidated damages clause reads:
If Contractor fails to achieve Substantial Completion of the Work on or before the Substantial Completion Date, Contractor must pay $2,500.00 per Day for each Day that Contractor fails to achieve Substantial Completion.
What it means: Every day past the Substantial Completion Date costs $2,500. That belongs in the schedule and the price, not in a contingency line.
None of these show up in a takeoff. All of them change what the job is worth.
What "evidence" means in an estimate
Estimating with evidence does not mean removing judgment. It means every judgment points to something you can show.
For each risk, an evidence-based estimate records four things:
- What the requirement says, quoted, with the clause and page.
- Who it falls on: you, a sub, or the owner.
- The decision: price it, qualify it if the solicitation allows, or carry it.
- Why: the one-line reasoning someone else can follow later.
A contingency line says "something might go wrong." An evidence-based line says "liquidated damages are $2,500 a day from the Substantial Completion Date, our schedule has 15 days of float on the critical path, and we have carried the rest." One of those can be defended at a bid review. The other cannot.
Why this is hard on a real bid desk
It is not a skill problem. Experienced estimators usually know where the risk is. The problem is time. A package like this one runs to 803 pages, 571 of them specifications, and the bid is due in a few weeks alongside four others.
So the reading gets rationed. The drawings get full attention. The contract and Division 01 get skimmed. The risk gets covered with a percentage. Sometimes that is enough. Sometimes it is the difference between a good job and the lowest bid on a bad one.
Where software helps, and where it doesn't
Software is good at the part people run out of time for: reading every page, finding every requirement, and showing where each one is. It should not make the decision. Whether to price, qualify or carry a risk is the estimator's call, and it should stay that way.
The test for any tool is simple. Can it show you the exact clause and page behind everything it tells you? If not, you are back to trusting a summary, which is just a different kind of guess.
Start with the contract
The quickest improvement most teams can make is to read the contract terms and Division 01 before pricing, not after. List the terms that carry money, decide on each one, and write down why.
That list does not just make the bid better. If you win, it becomes the project team's first look at the risks they now own.
Common questions
Why do bids on the same project vary so much?
Bidders read scope, schedule and contract risk differently. On one public package bid in March 2026, three bids ranged from $6,476,000 to $8,251,823, a spread of about 27%.
What is an evidence-based estimate?
One where every risk decision points to something you can show: the requirement quoted with its clause and page, who it falls on, whether you priced, qualified or carried it, and a line of reasoning someone else can follow.
Where should AI help in construction estimating?
In the reading: finding every requirement and term across hundreds of pages and showing exactly where each one is. The decisions on what to price, qualify or carry should stay with the estimator.
See the risk in your next bid, with the page behind it.
We read the contract, Division 01 and every addendum, and list the terms that carry money, each cited to the page, so your team decides with the evidence in front of it.