Articles · Cost estimating

How to Review a Contractor's Estimate Before You Sign

Seven checks to run on a contractor's proposal before you sign it: quantities against the drawings, missing line items, unit prices that don't hold up, hidden exclusions, contingency and who carries which risk.

Published 8 min readPr3cise

Lines of a takeoff sheet with one item flagged in gold

You have three bids on the table. The low one is 20% under the other two and somebody wants to sign this week. Reviewing a contractor's estimate isn't checking that the math adds up. It's working out what the document says, what it doesn't say, and who owns the risk of everything it doesn't say.

The same checks work whether you're an owner reading a GC's proposal, an architect or PM advising a client, or a general contractor leveling sub bids. The order is the same, and you can get through it in an afternoon if you know where to look.

What you're actually reviewing

An estimate isn't a price. It's a proposal for how risk gets divided. Every line says who pays if the work doesn't match the drawings. So the review starts with a question about the documents it was built from, not about the number.

AACE International's classification for building and general construction (RP 56R-08) puts figures on what to expect: a concept-stage Class 5 estimate typically lands between -20%/-30% and +30%/+50% of actual cost, while a Class 1 estimate built on a detailed takeoff of a finished set narrows to -3%/-5% and +3%/+10%. If someone hands you a Class 5 number and you're about to sign it as a lump sum, the number isn't the problem. What you're doing with it is.

This isn't about distrust either. Arcadis (2025) found that errors and omissions in contract documents were again the leading cause of construction disputes in North America in 2024. Most fights don't start with bad faith. They start with an incomplete document both parties signed.

Check 1. Quantities against the drawings

Go after the lines that carry the money

You don't have to rebuild the estimate. On almost any job, three to five divisions carry most of the cost: structure, envelope, MEP and finishes. Start there, and inside each one take the two or three largest line items.

Recompute it, don't read it

Pull the drawings and do the takeoff yourself. Hypothetical example: a 15,600 SF elevated slab at 10 inches thick is about 481 cubic yards of concrete. If the bid carries 430 CY, it's 51 CY short, or 10.6%. At a hypothetical $185 per cubic yard in place, that's $9,435, before the rebar, formwork and crane time that were priced off the same quantity.

Ask for the takeoff conventions alongside the number: whether openings are deducted, whether it measures to centerline or face, whether laps and waste are carried. Two correct takeoffs using different conventions produce different quantities, and that gap turns into an argument in the field later. An estimate you can check shows the quantity next to the line item, the way our sample estimates do.

Check 2. What isn't there

A missing line item isn't a discount. It's a deferred invoice. In a study of highway paving contracts, Bajari, Houghton and Tadelis (2014) estimated the cost of adapting incomplete contracts after award at 7.5% to 14% of the winning bid, far more than what bidders gained from private information or market power. What doesn't get taken off gets paid for later, at a worse price.

The usual absentees:

  • General conditions: superintendent, trailer, fencing, temporary power, security, dumpsters, cleanup.
  • Temporary works: shoring, scaffolding, bracing, hoisting and crane time.
  • Waste and overage factors, plus test and breakage material.
  • Protection of finished work, safety and site logistics.
  • Testing, special inspections, commissioning, closeout documents and as-builts.
  • Permit fees, utility connections and temporary services.
  • The seams between trades, where each sub assumes the other one carried it.

That last one moves the most money. Between the electrician and the low-voltage sub, between the framer and the drywall sub, there is always a strip nobody priced because both assumed it sat in the other scope.

Check 3. Unit prices that don't hold up

You can only judge a unit price if you know what's inside it. Ask for the buildup: material, labor with a production rate, equipment and the indirect percentage. A unit price with no buildup can't be compared to anything.

Hypothetical example of what to look for on a bid tab: three proposals for the same scope at $846,000, $902,000 and $694,000. The two high bids average $874,000, so the low one sits $180,000 under, 20.6% below. That spread is almost never efficiency. It's missing scope, an impossible production rate, or a material price with no quote behind it.

Public procurement has been proving this for decades. In Italy, Decarolis (2014) found that at least half the savings from awarding public works at the lowest price were lost to renegotiation afterward. Reviewing a Spanish high-speed rail program, the Court of Audit (2017) found works contracts awarded 23.3% below the tender price ended up 18.4% above their award value, and that 49% of the works contracts examined were modified, adding 11.6% in cost, almost entirely because of errors or omissions in the original design rather than unforeseen conditions.

Check 4. Read the exclusions first

The exclusions page gets read before the numbers, not after. That's where risk gets transferred, and it's usually written in language that reads as harmless:

  • “Unless otherwise noted” or “anything not expressly described”: an open-ended exclusion that fits anything.
  • “By others”: work somebody will have to buy separately.
  • “Assumes clear site access” or “assumes existing structure is sound”: assumptions that become change orders when they fail.
  • Requirements specific to the market you're building in: seismic bracing in one place, product approvals for a high-velocity hurricane zone in another, energy-code commissioning in a third.

Turn every exclusion into one question: if this happens, who pays? If the answer isn't in the document, you don't have a price yet.

Check 5. Contingency, allowances and escalation

How much contingency, and whose is it

Contingency isn't a cushion for short quantities. The U.S. General Services Administration (2024), which builds and renovates federal buildings, carries 7% on new capital construction and 10% on repair and alteration work, and says plainly that contingency should follow each project's level of risk. In the UK, HM Treasury's Green Book guidance (2003) tells public bodies to uplift early capital estimates for optimism bias by up to 24% on standard buildings and 51% on non-standard buildings including refurbishments, dropping toward 2-6% only as specific risks are identified and managed.

Hypothetical example of why it matters: a $4.2 million estimate with 5% contingency carries $210,000. If the $1.15 million structural package is 10% short on quantities, that's $115,000, or 55% of the contingency gone before the first differing site condition shows up.

Ask the three questions that define a contingency: who controls it, what can draw on it, and what happens to anything left at closeout.

Allowances are unpriced risk

An allowance is a blank with a placeholder number in it. It's legitimate when the scope genuinely isn't defined, but every one you sign is a future negotiation with no competition and the job already running. Count how many allowances the proposal carries and what they total: that percentage is the share of your contract that still has no price.

Escalation and bid validity

Check the bid validity date and what happens to material pricing if the job starts six months out. The AGC (2026) reported input prices for new nonresidential construction up 7.1% from June 2025 to June 2026, roughly double the 3.5% rise in contractors' bid prices, which suggests contractors were absorbing much of it. With no escalation clause, somebody is carrying that difference without having decided to.

Check 6. Who carries which risk

The contract type divides the risk before anybody breaks ground, and it's worth saying out loud during the review:

  • Lump sum: the contractor owns quantity and pricing risk unless the owner changes the scope. It only works on a defined set; otherwise you get change orders instead.
  • Unit price: the contractor owns the price, the owner owns the quantity. It needs disciplined field measurement and written measurement conventions.
  • Cost plus, with or without a GMP: the owner carries most of the cost risk in exchange for transparency and an earlier start.

When the split isn't clear, somebody else ends up deciding it. In HKA's database (2025) of more than 2,200 projects across 114 countries, sums in dispute averaged 33.4% of contract budgets and contractors' extension-of-time claims averaged 65.8% of the planned schedule.

Check 7. Make the schedule and general conditions agree

General conditions are a function of time, so you can test them with one multiplication. Hypothetical example: a 9-month schedule at $14,500 a month is $130,500. If the proposal carries $96,000 of general conditions, either the contractor plans to finish in 6.6 months or $34,500 is missing.

Both answers tell you something. If the real schedule is shorter than the contract schedule, ask how. If general conditions are short, you now know where the first change order is coming from.

Red flags and what they usually mean

Red flagWhat it usually meansWhat to ask for
One bid 20% under the restMissing scope or a takeoff errorThe buildup for the largest division and the quantities behind it
Round, repeating unit pricesA lump sum spread backward, with no buildupMaterial, labor with production rates, equipment and indirects
A long list of allowancesUnpriced risk to be settled mid-jobA unit price and a reference quantity for each one
No bid validity dateEscalation left openA validity period and an escalation clause or index
General conditions that don't match the scheduleThe bid schedule isn't the build scheduleThe schedule and the monthly general conditions rate
Broad, open-ended exclusionsRisk transferred to whoever signsA closed list of exclusions, item by item

The low bid is only the low bid if it covers the same scope as the others. Level the scopes before you compare the prices, or you're not choosing a contractor. You're choosing whoever left the most out.

How to level three bids without getting burned

Leveling means putting every proposal on one sheet, division by division, and adding back to each one whatever it needs to cover the same scope. The order at the bottom is usually not the order you started with.

  1. One list of divisions, yours, not each bidder's.
  2. One column per bid, with their numbers mapped onto your list.
  3. One row per exclusion, priced, added to whoever excluded it.
  4. One row for allowances, so you can see how much of each contract still has no price.
  5. A final comparison that includes schedule, general conditions and payment terms, not just the total.

If leveling turns up a large scope gap, don't settle it by email. Get the bidders on a call and have each one walk their quantities. Differences in convention show up in ten minutes and cost nothing to fix before signing. What happens when nobody does this is covered in the risks of an inaccurate construction estimate and in the most common ways to ruin a construction project.

How we work

Pr3cise is a remote estimating office working for owners, architects, general contractors and specialty contractors across the US, Spain and the UK. We do two things with the bids you receive: we take the work off independently from the drawings, so you have your own number to check against, and we review what you already have line by line, looking for scope gaps, unit prices that don't hold up and exclusions that quietly move risk onto you. The scope of that work is on our services page.

If you're deciding between proposals, or uneasy about a single one, send us the plans and the bid. And if you're also weighing whether to build that capability internally, we ran the numbers in in-house estimator vs outsourced estimating.

Frequently asked questions

The exclusions, before the numbers: that's where risk gets transferred. Then the three to five divisions that carry most of the cost, recomputing the quantities from the drawings instead of reading them. Finally, make sure the takeoff conventions are written down: two correct takeoffs using different conventions produce different quantities.

Ask for the buildup: material, labor with a production rate, equipment and the indirect percentage. A unit price with no buildup can't be compared to anything, because you don't know whether it carries waste, equipment or indirects. If one bid sits 20% below the average of the others, the usual explanation isn't efficiency. It's missing scope, an impossible production rate, or a material price with no quote behind it.

It depends on how well the scope is defined and how risky the job is, not on a fixed number. For reference, the U.S. General Services Administration carries 7% on new capital construction and 10% on repair and alteration work, and says contingency should follow each project's risk. At an early stage, HM Treasury's Green Book guidance tells UK public bodies to uplift capital estimates by up to 24% on standard buildings and 51% on non-standard ones for optimism bias, falling toward 2-6% only once specific risks are identified and managed. Always ask who controls the contingency, what can draw on it and what happens to anything left over.

Have a project to price?

Send us the plans and get back an itemized, bid-ready estimate with quantities and current market pricing.

Request an estimate