Most estimating disputes don't start with bad math. They start when a number built for one purpose gets treated as another. A budget check scribbled during a first meeting turns into the owner's "your number" six months later, and now you're defending a price you never meant to commit to.
The fix is knowing which of the types of construction estimates you're producing, what it's built from, and how it fails. There are five main types of construction estimates: order of magnitude, square foot, assembly, unit price, and detailed. They run from least to most design information, and only the detailed estimate should carry a firm price. Below, each one gets its inputs and how it fails, and a decision table at the end matches each type to project stage and what to put in writing.
Types of Construction Estimates Are a Promise About How Much You Know
Every estimate type is a statement about how much design information exists and how much risk the number carries. An early number built from a building type and a rough size is a different animal from a number built from a permit set and leveled sub quotes. Same project, different promise.
Labels vary by source. One firm's "conceptual estimate" is another's "schematic estimate," and some owners use "ROM," "budget," and "rough order of magnitude" interchangeably. Pick your own labels, define them in your estimating procedures, and use them consistently with owners and your team.
Why the same project gets three different numbers
*The following is a hypothetical, for illustration only.*
An owner asks you for a number on a 12,000 SF medical office building at three points. At the napkin-sketch stage, you give a wide range of $4.2M to $5.4M, based on two similar clinics you built. At schematic design, you price the major systems and land near $4.6M. At the full permit set, your detailed estimate comes in at $4.9M.
None of those numbers was wrong. Between stages, the owner added an imaging room that needs shielding, the finish schedule firmed up, and the civil drawings revealed a utility tie-in that nobody had sketched. Each change came from new information, which is exactly what each estimate type is supposed to absorb.
Why accuracy percentages don't travel well
You'll see tidy accuracy bands attached to estimate classes in many guides, such as "minus 30 to plus 50 percent" for early estimates. AACE International's estimate classification practices sort estimates into five classes, Class 5 (least defined) to Class 1 (most defined), and AACE itself warns that a class doesn't fix the accuracy: a clone of a job you've built carries far less risk than a first-of-its-kind building at the same design stage. That's why this guide uses qualitative confidence levels: low, moderate, high.
Set your own tolerances. Look at your last ten or so jobs, compare early estimates against final bid numbers, and write down the spread you actually see. That becomes your contingency policy, and it's worth more than a borrowed percentage.
Order-of-Magnitude Estimates: The Early Go/No-Go Number
An order of magnitude estimate is the rough, experience-based number you give before design exists. Its job is narrow: help an owner decide whether to keep going, and help you decide whether the job is worth chasing. It is not a budget, and it's definitely not a price.
What goes in and what comes out
Inputs are thin: building type, rough size, location, and your past comparable jobs. Maybe a sentence about quality level. That's about it.
The output should be a wide range, not a single figure. A single number from this stage implies precision you don't have. If you can't give a range, you're not ready to give a number.
How it goes wrong
The owner remembers the low end of the range. Always. They repeat it to their lender, their board, or their partners, and by the time drawings arrive, the low end is the "budget."
Put three things in writing every time: the range, the assumptions behind it, and the date. A two-line email works. "Based on two comparable clinics, no site information, no design, we'd expect $4.2M–$5.4M as of this date. This is not a price." That sentence is your defense later.
Square Foot Estimates: Fast, and Only as Good as Your Comparables
A square foot estimate multiplies a cost per square foot by the building area, then adjusts. It's a parametric method, meaning it relies on a relationship between cost and a single driver. It's fast, which is why owners and developers love it, and it's only as honest as the comparable jobs behind it.
Pull your comparables from your own completed work, not a published average. Match on building type, size band, structure, and delivery method. Three to five close matches beat fifteen loose ones.
The adjustments that matter most are site conditions, finish level, market timing, and complexity. A flat site with utilities at the curb and a hillside lot with rock are not the same $/SF, even if the building is identical. Escalate older jobs to today's dollars with a method you can explain.
Worked example: a hypothetical square-foot check
*Everything below is illustrative. The $/SF values are made up for this exercise and are not market rates.*
Assumptions: an 8,000 SF building. Three comparable past jobs from your own records came in at $310, $335, and $360 per SF (already escalated to today). Your adjustments are +3% for a more difficult site and +5% for a higher finish level. Contingency policy for this stage is 10%.
Steps:
- Combined adjustment: 1.03 × 1.05 = 1.0815.
- Midpoint comparable: $335 × 1.0815 = $362.30/SF. At 8,000 SF, that's $2,898,420.
- Low comparable: $310 × 1.0815 = $335.27/SF, or about $2,682,120.
- High comparable: $360 × 1.0815 = $389.34/SF, or about $3,114,720.
- Contingency on the midpoint: 10% × $2,898,420 = $289,842, for a midpoint with contingency of about $3,188,260.
Output: a range of roughly $2.68M to $3.11M before contingency, with a working midpoint near $2.90M. State what's excluded, such as land, design fees, owner equipment, and permits if you don't carry them.
To verify it, run the same math against two of your past jobs and see how far it lands from what they actually cost. If it's off by a lot, your comparables or adjustments need work. A single-driver method can't capture everything, and it needs professional judgment on anything unusual.
When to stop using it
Stop when your comparables stop matching. Unusual structures, heavy site work, tight renovations, and scopes with owner-furnished equipment all break the single-driver assumption. A $/SF number on a gut renovation can be off by more than the contingency you'd carry.
Once design firms up and the numbers need a real layout, move to a proper template. The cost estimate template post covers how to structure it.
Assembly and Unit Price Estimates: The Middle of the Road
Once schematic drawings exist, you can price more than a single $/SF. Two middle-ground methods fit here: assembly and unit price. Both are among the more useful types of cost estimates in construction because they let you move faster than a full takeoff while still reflecting the actual design.
Assembly estimates for design-phase pricing
An assembly estimate prices building systems as units: an exterior wall type, a roof system, a slab on grade. Each assembly bundles its components (framing, sheathing, insulation, finish) into a cost per SF of wall or roof. You measure the system area, apply the assembly cost, and move on.
This is how you give a design team useful feedback without a full takeoff. If the architect is choosing between two wall types, you can show the cost delta in an afternoon. Build your assemblies from your own past jobs so the numbers reflect how you actually build.
Unit price estimates and where they show up in contracts
A unit price estimate prices installed work per unit: cost per cubic yard of concrete, per linear foot of pipe, per cubic yard of excavation. You multiply by quantity. It fits scopes where quantities may change after the contract is signed, like earthwork or rock excavation.
Because quantities can move, contract language matters. How are quantities measured, who measures them, and how is the overage or underrun paid? Those answers belong in the contract documents, and counsel should review the language.
How CSI divisions keep these organized
Assemblies and unit lines get messy fast without a consistent structure. Organizing them by CSI division keeps your price book, estimates, and sub scopes aligned. The CSI divisions field guide shows how to structure it.
Detailed Estimates: Quantity Takeoff, Pricing, and the Bid Number
A detailed estimate is built bottom-up from a plan-based quantity takeoff. You measure every scope item from the drawings, price each one, add subcontractor pricing, general conditions, overhead, contingency, and profit, then roll it all up. It's the most labor-intensive of the types of construction estimates, and it's the only one most contractors should sign a firm price on.
What a defensible detailed estimate contains
Every quantity should trace to a sheet. Every price should trace to a source: your company price book, a past job, or a current supplier quote. Subcontractor quotes should be leveled so you're comparing the same scope across bidders.
If you can't answer "where did this line come from?" in ten seconds, a reviewer can't either. For the mechanics, see how to do a construction takeoff and estimating a project from scratch.
Where subcontractor scope gaps hide
*Hypothetical example, for illustration only.*
Say you get two electrical bids on a job. Bid A is $184,000. Bid B is $211,600. That's about a 15% spread, and the temptation is to take the low one and move on.
Leveling shows A excludes the fire alarm system and temporary power. Assume, for this example, those are worth $14,000 and $6,500. Add them back and A becomes $204,500, a gap of about 3.5% against B. The real difference was scope, not price.
That's why bid leveling happens before the number goes into your estimate. Line the exclusions up first and scope differences surface before they turn into change orders.
Software and spreadsheets for the detailed stage
Detailed estimates get built in Excel, PlanSwift, STACK, Autodesk ProEst, Struvia, and a handful of other tools. What matters is whether the tool matches how your team works. The construction estimating software rankings and the Excel estimating playbook go deeper than we will here.
One plain contrast: takeoff-only tools stop at quantities, and you price the job somewhere else. Check whether a tool carries through to a priced estimate, and check current vendor terms before you buy anything.
Estimate Type Decision Table: Match the Number to the Moment
Here's the field-use output: one table that maps each estimate type to project stage, inputs, typical use, main failure mode, and what to put in writing. Treat it as a starting framework you can adapt.
| Estimate type | Project stage | Inputs needed | Typical use | Main failure mode | Put in writing |
|---|---|---|---|---|---|
| Order of magnitude | Pre-design, concept | Building type, rough size, location, comparable jobs | Go/no-go, early feasibility | Owner remembers the low end and treats it as a budget | Range, assumptions, date, "not a price" |
| Square foot | Early concept to early schematic | Area, $/SF from your own comparables, adjustments | Early budgeting, option comparison | Comparables don't match the real building or site | Comparable jobs used, adjustments, exclusions |
| Assembly | Schematic to design development | System quantities, assembly costs | Design-phase feedback, value engineering | Assemblies outdated or missing a component | Assembly list, drawing version, allowances |
| Unit price | Design development to bid, or contract unit pricing | Quantities by unit, unit costs | Scopes with changeable quantities | Unclear measurement and payment rules | Unit definitions, quantity basis, measurement method |
| Detailed | Permit or bid documents | Full takeoff, priced lines, leveled sub quotes | Firm bid, contract price | Scope gaps, missed addenda, unpriced items | Drawings and addenda used, exclusions, validity period |
How to read and adapt the table
Add your own contingency policy to each row. Early estimates carry more, detailed ones less, and the specific percentages should come from your own history, not from this article.
Define the point at which an estimate needs a second reviewer, too. That might be any number above a dollar threshold, any estimate for a new building type, or any bid where you're replacing a key estimator. The table is a framework, not an industry standard, and it doesn't replace judgment on unusual projects.
Bid estimate vs budget estimate: the line to draw
The bid estimate vs budget estimate distinction comes down to commitment. A budget estimate guides an owner's decision. A bid estimate commits you to a price.
Before any number leaves your office, label it. Use this checklist:
- Type: order of magnitude, square foot, assembly, unit price, or detailed
- Date: when the number was built
- Drawings used: sheet set and revision, including addenda
- Exclusions: what's not in the number
- Validity period: how long the number holds
If a number doesn't carry those five labels, it will get repeated as something it isn't.
Moving an Estimate Up the Ladder Without Starting Over
Estimates should get refined as drawings develop, not rebuilt each time. The goal is that every stage reuses the work of the one before it.
Keep a one-page assumptions log
Record every allowance, exclusion, and comparable job you used. Add the date and the drawing version. When the next set arrives, you can walk down the log and swap each assumption for real data.
Replace allowances with quoted prices as soon as quotes are available. Track what changed between versions, and why. When the owner asks why the number moved $200,000, you'll have the answer in one page instead of a week of digging.
Carry the same template from rough to final
Using one Excel template across all stages keeps line items consistent, so version-to-version comparisons are apples to apples. A rough estimate with the same structure as your final means you're filling in detail, not reorganizing.
Struvia delivers takeoffs and estimates in your own Excel template with your formulas intact, and can price from your company price book. That means a takeoff in hours, not days, in a format your team already knows how to review. The Struvia team can set up the template with you or you can do it yourself.
Frequently Asked Questions
What are the main types of construction estimates?
This article covers five: order of magnitude, square foot, assembly, unit price, and detailed. Some sources use different labels or split them into more classes, so match terms with your owners and team. The decision table above maps each one to stage, inputs, and failure mode.
What is the difference between a conceptual estimate and a detailed estimate?
A conceptual estimate is built from limited design information, usually with a parametric method like cost per square foot. A detailed estimate is built from a full plan-based takeoff with priced line items. The conceptual estimate vs detailed estimate difference is also about commitment: the first supports a decision, the second supports a firm price.
Which type of estimate should I use for a bid?
Use a detailed estimate built from a takeoff and leveled subcontractor quotes. The exceptions are bid forms that ask for unit prices or allowances, where you price those items per unit or as stated. Even then, the rest of your number should come from a detailed build-up.
How do I price a project when I only have a sketch?
Give an order-of-magnitude or square-foot range based on your own comparable jobs. State your assumptions, the range, and the date, and say plainly it isn't a price. Resist the urge to narrow the range to please the owner.
What is a unit price estimate in construction?
It's the cost per installed unit multiplied by the quantity, such as dollars per cubic yard of excavation. It fits scopes where quantities may change after award. Define how quantities get measured and paid in the contract so there's no dispute later.
Can I use a spreadsheet for every estimate type?
In principle, yes. Our free estimating spreadsheet and cost estimate template posts show how to set one up. Once your volume grows, manual takeoff tends to become the bottleneck, and that's when tooling starts to matter.
Pick the Estimate That Matches What You Know
The right estimate type depends on how much design information you have. Use a range early, systems pricing in the middle, and a detailed build-up before you sign anything firm. Label every number so it can't be mistaken for something else. If you want to see how the detailed stage works with less manual effort, try Struvia free on your next plan set and get a priced estimate back in your own Excel template. Struvia is free to try: you can run a few real plan sets before you pay anything, and no card is needed. Rates are on the pricing page, or book a demo to walk through a job.
*Reviewed by Weston Burnett, Co-Founder and CTO of Struvia.*
