How to Do a Division 8 Takeoff for a Commercial Bid

A Division 8 takeoff is not a matter of counting doors off a schedule and pricing what's listed. It's a reconciliation job, pulling door schedules, frame elevations, hardware sets, floor plans, and the 087100 spec into a single, coherent picture of every opening in the building, because none of those documents is complete on its own. Missing the reconciliation step makes the bid not just imprecise but short, in ways that become visible only at submittals or, worse, once the walls are framed.

Why Division 8's takeoff differs from every other trade

Division 8 carries the formal title "Openings" inside CSI MasterFormat, and it covers doors, frames, hardware, and everything else that fills a hole in a wall throughout a commercial building. MasterFormat itself is the shared filing system the industry runs on: a standardized way of organizing construction information so architects, engineers, contractors, and suppliers are all working from the same numbering scheme. An estimator working Division 8 will run into a handful of sections over and over. Section 08 11 00 covers metal doors and frames, hollow metal work, gauge, construction type, hardware prep, fire ratings, and finish. Section 08 14 00 covers wood doors, both flush and stile-and-rail, along with core type, face veneer, fire rating, and factory finish. And 08 71 00 covers door hardware, hinges, locksets, closers, exit devices, which is the section that actually drives the hardware schedule.

What makes this trade different from, say, a drywall or flooring takeoff is that no single document tells the whole story. Every opening ties together three separate document types at once, the schedule, the plan, and the spec, and none of the three stands alone as complete. The hardware schedule, usually an exhibit attached to 08 71 00, organizes doors into hardware sets under strict rules: labeled and non-labeled openings can't share a set, and doors of different sizes or materials can't share one either. Those rules matter because a takeoff process that reads documents one at a time, rather than cross-checking them, will build errors into the estimate before pricing even starts. And the scale compounds the problem. Estimators are working against document sets that can run 200–400 pages or more, with door and hardware details scattered throughout with no single consolidated location constructionbids.ai rockettakeoffs.com.

The documents a Division 8 estimator must assemble before touching a quantity

Nobody should start counting quantities before assembling the full document set. The first move on any Division 8 bid is a complete review of the drawings, the specs, the addenda, and any supplemental exhibits, because that review defines scope, exposes risk in the documents, and keeps the estimator from re-working the same opening three times later.

Five documents need to be on the table together, not reviewed in sequence. The door schedule is the master list, giving door number, location, material, dimension, fire rating, and handing for every opening. Frame elevations and door type drawings fill in what the schedule abbreviates: frame profile, lite kits, louver cutouts, the physical shape of the thing. Floor plans locate each opening in the building, establish handing in context, and often reveal openings the schedule never picked up. Partition and wall type schedules determine what the frame is actually being set into, which drives anchor selection and frame prep. And the 08 71 00 spec, along with its hardware schedule, defines every hardware product, grade, finish, and manufacturer, forming the contractual basis for the entire hardware package.

That 08 71 00 section follows a predictable three-part structure: Part 1 covers administration, submittals, and quality assurance; Part 2 covers products, including the hardware descriptions, grades, finishes, and manufacturers that actually get priced; Part 3 covers execution, installation, adjusting, cleaning. Pricing lives almost entirely in Part 2, but the compliance obligations run through all three, so skipping Parts 1 and 3 because "that's not where the numbers are" is how submittal requirements get missed securityparts.com. Addenda sit outside this five-document core but bind the bid just the same, whether or not anyone opened the file before submission. The discipline that separates a clean bid from a scramble is logging each addendum against the base set the day it's issued, noting immediately if it shifts a quantity, swaps a spec section, or adds scope, rather than letting addenda pile up into a stack reviewed the night before the bid is due.

Extracting the door schedule

The door schedule is the spine of the whole exercise. Every other document either feeds into it or corrects something it got wrong. For each opening, the extraction needs to capture the door number and location, the door and frame materials, the dimensions (width by height by thickness), the fire rating if one applies, the handing, and the hardware set designation that ties the opening to its hardware group.

One extraction problem occurs constantly: the door schedule usually lives on a drawing sheet, but a supplemental hardware schedule often lives in the spec book, and the two don't always agree. Catching that mismatch at this stage, before pricing starts, is what keeps a discrepancy from turning into a repricing exercise weeks later. The discipline worth building here is standardizing the extracted data, opening number to hardware set to component list, into one consistent format before moving on to plan verification. That structure is what separates an actual takeoff from a running tally that happens to have numbers in it. Nothing gets priced and no hardware gets assigned at this stage. The only output is a clean, complete list of openings with every field populated, row by row.

Reconciling the schedule against floor plans to catch what the schedule misses

The door schedule is an architect's document, and architects revise drawings faster than they revise schedules. The floor plan, by contrast, is the legal depiction of the building, so any conflict between the two is the estimator's problem to sort out before bid day, not something to discover after the contract is signed.

Every opening on the floor plan needs to be checked against a short list of things the schedule can get wrong. Is the opening physically present, and does it match the door number tag it's supposed to carry? What's the configuration, single, pair, equal pair, unequal pair, since that changes hardware quantities for astragals, coordinators, and flush bolts? Does the handing shown on the plan match the handing noted on the schedule, since schedules get handing wrong more often than most estimators expect? What wall type sits at the opening, since that determines anchor type and frame profile? And, critically, are there openings on the plan carrying no schedule tag? Those untagged doors are the ones that blow up a bid weeks after it's submitted.

The rule that matters here: count every door, frame, and configuration type straight off the plan, not off the schedule's total. The schedule is a reference document. The plan is ground truth.

Assigning hardware sets and sizing accessory items to the opening

Hardware set assignment means matching each opening's designation from the door schedule to its corresponding group in the 08 71 00 spec. That spec is the contractual basis for the hardware package, so any deviation from it needs a submittal and an approval, not a field decision.

Hardware groups are built around openings that function the same way, and the grouping rules are stricter than they look at first glance. Labeled and non-labeled openings can never share a set. Doors of different sizes or different materials can't share a set either, even when they look functionally identical on the plan.

Accessory sizing is where a lot of otherwise careful takeoffs go wrong, mostly because accessories get treated as a single catalog item instead of something sized to the specific opening. Kick plates are sized by leaf width, typically the door width minus two inches, plus a specified height, an item sized to the specific opening. Thresholds are sized by opening width, and the exterior or interior application changes both the profile and the material. Astragals apply only to paired doors, and their width is set by the inactive leaf's dimension, not a standard size pulled from a catalog page. Treating any of these as one-size-fits-all is how ordering errors and field shortages happen. Handing needs to be confirmed at this stage too, since it affects lever function, closer arm type, and frame prep, and a reversed hand caught at submittal instead of at takeoff means repricing hardware that's already been counted once.

Hidden scope in specification requirements: fire ratings, ADA, ANSI/BHMA grades, and divided-scope items

Fire-rated assemblies carry their own strict compliance chain. They require UL-listed hardware that complies with NFPA 80, and the fire rating on the schedule has to be matched by every single component in that opening's hardware set. A closer or a hinge that isn't listed, sitting on a labeled opening, is a rejection at inspection, full stop, no argument to be had about it.

Accessibility requirements under ADA and ICC A117.1 apply to every opening on an accessible route. Hardware has to operate with one hand and without tight grasping or twisting. Operating force can't exceed 5 pounds of force. Mounting height has to fall between 34 and 48 inches above the finished floor. None of this is a nice-to-have upgrade; it's code, and it constrains hardware selection on any opening that sits on an accessible path.

Grading adds another layer. BHMA is the only ANSI-accredited body for builders' hardware performance standards, and Grade 1 is the baseline that commercial specs are written around. Under A156.2, a Grade 1 cylindrical lock has to survive one million cycles securityparts.com. Under A156.3, a Grade 1 exit device has to survive 500,000 cycles under preload securityparts.com. Those numbers aren't marketing claims, they're the test thresholds a product has to clear to carry the grade.

Edition years are a quieter trap. Standards get reviewed on a five-year cycle, and a spec that cites a standard number without naming the edition year hasn't actually specified which version applies. That gap lets an older, technically certified product satisfy the letter of the spec while missing what the architect clearly intended. Before pricing any alternate, the edition year cited needs to be confirmed, not assumed.

Divided scope is probably the most commonly missed cost in the whole division. Division 8 might specify the electrified hardware itself, while Division 28, Electronic Safety and Security, carries the access control system or the power supply behind it, and an estimator has to know exactly where that line falls before the bid goes out. Frame anchors, fire-rated glazing, vision panels, louvers, protection plates, coordinators, door bottoms, each of these touches both the function and the cost of an opening, and each has to be checked against the spec individually before the bid closes.

Where drawing-spec conflicts hide before bid day

Post-award disputes rarely come from some dramatic last-minute change order. They come from mismatches between the drawing set and the spec book that nobody caught during takeoff. HKA's 2025 CRUX Insight report, which analyzed more than 2,200 disputed projects worldwide, found that change in scope was the single leading cause of construction claims globally, and drawing-spec conflicts sit inside that category as a named, recurring contributor constructionbids.ai.

In Division 8 specifically, the conflicts tend to follow a handful of recognizable patterns. A door schedule lists a fire rating that the frame elevation drawing never shows, or the reverse. A floor plan shows a pair of doors where the schedule lists a single. A hardware set in the spec references a product that's since been discontinued or substituted by addendum, but nobody went back and reconciled that addendum against the schedule. A partition schedule calls for CMU construction while the frame anchor spec assumes stud framing. The anchors ordered won't work in the wall that actually gets built⟧c29⟧.

The dollar exposure on these misses scales with the size of the job. The only dependable way to catch these before bid day is to read all five document types against each opening simultaneously, not one document at a time in sequence, because sequential review is exactly the process that lets these conflicts slip through. Addenda that shift quantities, swap out a spec section, or add scope need to get reconciled against the base opening list as part of that same simultaneous pass, not treated as a separate stack reviewed only if there's time left before submission. A 3% measurement error on a $10M project translates to $300,000 in exposure, a figure worth framing against the size of a typical Division 8 package to show what a handful of missed conflicts costs constructionbids.ai.

Institutional jobs and owner-standard hardware: when the spec is a starting point, not the finish line

On a standard commercial bid, the 08 71 00 spec is the full contractual definition of what gets supplied, and the estimator's job is simply to extract it and price it accurately. Institutional work runs differently. On hospitals, universities, and government facilities, the project-specific spec often takes a back seat to the owner's hardware standard, a master template the owner maintains across its entire building portfolio, not just this one job.

That changes the estimator's task in a real way. Instead of just extracting what the project spec says, the estimator has to compose hardware sets that comply with the owner's standard, which can override manufacturer substitutions the project spec would otherwise permit. Owner standards frequently lock in specific product lines, finish standards, and keying systems that have to stay consistent not just across every opening on this project, but across every project the owner runs. A spec clause reading "or approved equal" can still be functionally locked to one specific product by the owner's standard, and an alternate that satisfies ANSI/BHMA Grade 1 on paper will still get rejected if it isn't on the owner's approved list.

Missing that owner-standard layer doesn't produce a bid that's slightly off. It produces a bid for a different scope of work than the one the owner actually intends to buy, and fixing that after award turns into a change-order fight rather than a simple substitution request. The only reliable way to know whether a project carries owner standards is to read the front-end spec sections and any referenced owner design-standards documents, not just 08 71 00.

In 2023, 26% of construction firms reported using AI somewhere in their workflow; by 2025 that figure had climbed to 37%, a jump of 42% year over year, and adoption among the top 100 general contractors crossed 60% sometime during 2025 bidicontracting.com nedesestimating.com constructionbids.ai.

On Division 8 documents specifically, these tools work by combining a few distinct capabilities. Computer vision and machine learning models detect and quantify building components straight from digital drawings, recognizing door swings on a floor plan and tying them back to the opening tags they belong to, including flagging rough-opening dimensions that need a second look. Named entity recognition parses the spec documents themselves and cross-references what it finds against plan callouts, which lets hardware set assignment happen automatically instead of by hand. The real advantage appears in reconciliation. Cross-checking multiple document types against each other is the hardest, most error-prone part of doing this manually, and it's exactly the part AI tools handle in parallel rather than one document at a time.

The efficiency numbers back this up. General-purpose takeoff tools built for broad building quantities don't automatically understand the particular grammar of Division 8. The reconciliation logic still has to be built for openings specifically, using logic distinct from a tool designed to count square footage of drywall. An AGC Technology Survey cited by constructionbids.ai found that digital takeoff software reduces quantity surveying time by 62% and measurement errors by 84%, while nedesestimating.com reports that AI tools cut takeoff time by 70% or more and push accuracy above 95% on standard commercial and residential plans rockettakeoffs.com. SOURCE PAGES (what the pages behind the outline's links say).