Why Do Architectural Drawings Contradict Themselves?

November 2, 2026

Short answer: Architectural drawing inconsistencies happen because the same wall, door or ceiling is drawn on a plan, an enlarged plan, an elevation, a section, a detail and a schedule, and a change made on one view does not always reach the others. Across 523 adjudicated conflicts in 19 real plan sets, an architectural sheet contradicting another architectural sheet accounts for 10.7% of everything found, almost as many as the most common cross-discipline pair.

The floor plan tags the corridor door 20-minute. The door schedule lists it 90-minute. The enlarged unit plan shows a 3′-0″ opening where the floor plan dimensions 2′-10″. The wall type tag on A-201 points to type 4B, and the wall type legend stops at 4A. Section 3 on A-401 calls for a 9′-0″ ceiling that the reflected ceiling plan sets at 8′-6″. Nobody else’s sheets are involved. The architectural set disagrees with itself.

Most talk about coordination is about trades clashing: duct against beam, light against diffuser. That is real, but it is not the whole job. In our benchmark, about one conflict in four is a sheet contradicting another sheet of its own discipline, and architectural is by far the biggest source of those.

This is where architectural sets break internally, why it happens on sets drawn by good firms, and how to check an architectural set against itself on a PDF before pricing or GMP.

How common are architectural drawing inconsistencies?

Across 523 adjudicated coordination conflicts in 19 real plan sets totalling 5,259 sheets, 56 are an architectural sheet contradicting another architectural sheet. That is 10.7% of everything found. The single most common cross-discipline pair, Architectural against Electrical, is 11.1%. If internal architectural conflicts were ranked as a pair, they would sit second, ahead of Architectural against Mechanical at 9.4%.

Architectural shows up in about half of all the conflicts in the study, against every other discipline and against itself. It is the hub of the set. The next largest internal source is electrical, at 33 conflicts or 6.3%, usually a panel schedule that disagrees with the power plan it describes. Every other discipline is at 2% or below.

Intra-discipline conflicts as a whole come to 139 of 523, about one in four. A review that only lays one trade over another misses a quarter of the problems before it starts.

Why architectural sets disagree with themselves

An architectural set describes one building many times over. A single corridor wall appears on the floor plan, the enlarged plan, the life safety plan, the reflected ceiling plan, a building section, a wall section, a wall type legend and maybe a detail. A single door appears on the plan, the enlarged plan, the elevation, the door schedule and the hardware groups. Each view is a separate drawing of the same object.

The Uniform Drawing System in the United States National CAD Standard lays out how a set is organized into sheet types and how sheets reference each other, and a disciplined office follows something like it. What no standard does is keep the views in agreement. When a door gets a rating upgrade late in design, someone has to change the plan tag, the schedule, the hardware group and the life safety plan. Miss one and the set now says two things.

Three things make it worse. Architectural sets are the largest discipline by sheet count, so there are more views to keep in step. They are drawn by several people at once, often with the enlarged plans and details handled by someone other than the person running the floor plans. And they change the most, because the architect is the one fielding owner changes, code comments and the other consultants’ coordination. Each round of revisions clouds the change on the sheet where it was made, and the sheets where it should also have been made carry no cloud at all.

The eight places architectural drawing inconsistencies show up

They look different on the job. On paper they are the same thing: two views of one object that no longer agree.

1. Door tags against the door schedule

Size, swing, type, frame, hardware group and fire rating all live in two places: the tag on the plan and the row on the schedule. Tags that point to rows that do not exist, rows with no tag on any plan, and ratings that differ between the two are the most common internal conflict in an architectural set. The rating mismatch is the expensive one, because the door gets ordered from the schedule and inspected against the life safety plan.

2. Wall type tags against the wall type legend

The plan tags a wall with a type. The legend defines the type: stud size, layers of board, insulation, rating and the tested assembly it relies on. Tags for types that are not in the legend, legends that changed a type without the plans following, and a rated corridor wall tagged with a non-rated type all leave the drywall sub pricing one wall and the inspector looking for another.

3. The life safety plan against the floor plans

The life safety or code plan shows rated walls, rated doors, exit paths and occupant loads. The floor plans are supposed to build what it shows. When a wall rated on the life safety plan is tagged as a plain partition on the floor plan, or an exit door swings the wrong way on one of them, the set disagrees about something an inspector will check.

4. Enlarged plans against the floor plans

Toilet rooms, kitchens, stairs and, on multifamily work, every unit type get an enlarged plan. The floor plan shows the same rooms at a smaller scale. Dimensions, door locations, fixture counts and wall positions drift apart as one gets revised and the other does not. On a multifamily set, a unit plan revised once can disagree with every floor it repeats on, and a mirrored unit is a second chance for the same mistake.

5. Ceiling heights across the RCP, sections and interior elevations

The reflected ceiling plan gives a height. The building section, the wall section and the interior elevations each show the ceiling too. When they disagree, the framer, the drywall sub and every trade working above the ceiling are pricing a different plenum.

6. Detail and section callouts that point to the wrong place

A callout bubble says detail 7 on A-502. Detail 7 on A-502 is a different condition, or there is no detail 7, or A-502 was renumbered in the last issue. Broken references are small one at a time and add up fast across a large set, and each one is an RFI the architect has to answer before the work can be priced.

7. Elevations against plans and the window schedule

The exterior elevation shows a window count, sizes and head heights. The floor plan shows openings in the wall. The window or storefront schedule lists the units. When a window moves on the plan and the elevation still shows the old layout, or the schedule lists a size neither drawing shows, the glazing sub has three answers to choose from.

8. Finish schedules, room names and room numbers

The finish schedule keys finishes by room number. The plans, the finish plan and the interior elevations name and number the rooms. Renumbered rooms, rooms missing from the schedule and finish codes with no legend entry are low severity one at a time, and they are where scope goes missing at bid.

Why these cost more than they look

Processing a single RFI runs about $1,080 in loaded review time across the builder, the architect and the engineers, and projects average roughly ten RFIs per $1 million of construction value. A conflict found in the field instead of on paper averages around $1,500.

Internal architectural conflicts hit at two moments. At bid, they turn into scope gaps and allowances, because the estimator has to pick one of two answers and the sub picks the other. In the field, they turn into wrong material on site: a door frame ordered from the schedule that does not match the rated opening on the life safety plan, or a ceiling framed at the height on the section while the ductwork was coordinated to the RCP.

Found on paper, each is one RFI and a revised sheet. The difference is whether anyone read the schedules next to the plans before anything was bought.

How to check an architectural set against itself on a PDF

You do not need a model for any of this. Most of it is a schedule walked against the plans, and the rest is two views of the same room side by side.

  • 1. Confirm every architectural sheet is on the current issue. Check the revision block on every A sheet against the sheet index. A sheet left on an older issue explains a whole cluster of conflicts on its own.
  • 2. Walk the door schedule against the plans, both ways. Every tag on a plan should have a row, every row should have a tag, and the rating on the row should match the rating the life safety plan requires at that opening.
  • 3. Walk the wall type tags against the legend. Every tag should resolve to a type in the legend, and every rated wall on the life safety plan should carry a rated type on the floor plan.
  • 4. Lay each enlarged plan over its floor plan. Check dimensions, door locations and wall positions. On multifamily work, check each unit type against every floor it appears on, and check mirrored units separately.
  • 5. Compare ceiling heights across the RCP, sections and interior elevations. One height per room. Any room with two is a finding.
  • 6. Follow every detail and section callout. Each bubble should land on a detail that exists, on the sheet it names, showing the condition it was called for.
  • 7. Check elevations against plans and the window schedule. Window count, size and location should agree across all three.
  • 8. Match room numbers across the plans and the finish schedule. Every room on the plans should have a finish row, and every finish code should have a legend entry.
  • 9. Write every finding with both sheet numbers. “A-201 tags door 214 as 20-minute where A-601 door schedule lists 214 as 90-minute” gets answered in one reply. “Check the door ratings” does not.

Why clash detection and discipline reviews miss these

3D clash detection compares geometry from different models. It does not compare a door schedule to a door tag, a callout bubble to the detail it names, or a life safety rating to a wall type. A model built from an inconsistent set inherits whichever answer the modeller picked, and the other answer disappears without anyone deciding it should. On most projects pricing happens on 2D PDFs before any model exists anyway, which is why clash detection without BIM matters at bid.

Discipline reviews have a different blind spot. The architect’s own QA usually checks one sheet at a time, often by the person who drew it. The consultants check their own sheets. Nobody is paid to read the door schedule against every plan tag in a 400 sheet set. That is the work a construction document review checklist is meant to force, and it is worth doing before you review drawings before GMP.

How flikt.ai/ handles it

flikt.ai/ reads the whole set, every sheet against every other sheet including sheets of the same discipline, plus the schedules and any specifications, and returns a report of the contradictions with page references and a severity rating on every finding, the same day. It classifies findings across 16 disciplines and works directly from 2D PDFs, so there is no model to build and nothing to install.

Pricing is published, and every listed price is a ceiling: checkout prices from the pages you actually upload. A Coordination Review is $279 for a set up to 80 pages and $399 up to 120 pages. A full CD set of around 250 pages is $1,049, and a multifamily or mixed-use set of around 800 pages is $3,799. After the first review, Project Coverage runs monthly on that project and covers every re-review while the job is live. Teams with several jobs open tend to land on a Review Desk at $599 a month. Current tiers are on the pricing page, and you can get a free preview on your own set first.

Reports are recommended as a supplement to professional review, not a replacement for it.

Frequently asked questions

Why do architectural drawings contradict each other?

Because the same wall, door or ceiling is drawn on several views: plan, enlarged plan, elevation, section, detail and schedule. In a 2D set those views are separate drawings, and a change made on one does not always reach the others. Late revisions, several drafters working at once and the sheer number of architectural sheets all add to it.

How common are architectural drawing inconsistencies?

Across 523 adjudicated coordination conflicts in 19 real plan sets totalling 5,259 sheets, 56 were an architectural sheet contradicting another architectural sheet, or 10.7%. That is close to the most common cross-discipline pair, Architectural against Electrical, at 11.1%. About one conflict in four across the study was within a single discipline.

What is the most common conflict within an architectural set?

Door tags that disagree with the door schedule: a size, a type, a hardware group or a fire rating that differs between the plan and the schedule, or a tag with no row. Wall type tags that do not match the wall type legend, and enlarged plans that have drifted from the floor plans, are close behind.

How do I check a door schedule against the floor plans?

Go both ways. For every door tag on the plans, find its row on the schedule and compare size, type, frame, hardware and rating. Then for every row on the schedule, find the tag on a plan. Finally, check the rating on each row against what the life safety plan requires at that opening. Write each mismatch with both sheet numbers.

Does 3D clash detection find architectural inconsistencies?

Mostly no. Clash detection compares geometry between models. It does not compare a schedule to a tag, a callout to the detail it points at, or a fire rating on one sheet to a wall type on another. A model built from an inconsistent set simply takes one of the two answers.

Who is responsible for internal conflicts in the architectural drawings?

The architect of record owns the architectural set and is responsible for its internal consistency under most design contracts. In practice the builder is the one who finds out, at bid or in the field, so it pays to check before pricing and send the conflicts back as RFIs with both sheet numbers.

Can you find architectural drawing inconsistencies without a BIM model?

Yes. Almost all of them are a schedule against a plan or one view against another, which is a PDF check. Walk the door schedule and wall types against the plans, lay enlarged plans over floor plans, follow every callout and compare ceiling heights. Automated review tools like flikt.ai/ run the same cross-check across every sheet of a 2D PDF set, with no model to build.

Free field guide

The pre-IFC coordination checklist

50 checks before you issue for construction, ordered by what actually goes wrong — built from 1,516 findings across 34 real plan sets. Comes as an assignable tracker (Excel) with status, owner and due date, plus a printable PDF. No call, no card.

Name
← Back to Resources
1