What is plan check, and why does it take so long?

The job a plan checker is actually doing#

A plan checker is not evaluating whether your house is well designed. That is not the assignment. The assignment is to verify that the documents in front of them demonstrate compliance with the codes and standards the jurisdiction has adopted, and to write down everywhere the documents fail to do so.

The distinction matters because it explains most of what frustrates owners about the process. A reviewer cannot approve something they can reasonably infer but cannot find drawn. If a detail is missing, the correct outcome is a correction, even when everyone involved knows how the condition will be built. Plan check is a documentation review, and the document is the deliverable.

The second thing to understand is that plan check is not one person. A submittal is typically split across several reviewers working in parallel, each with their own scope, their own comments and their own idea of what a complete sheet looks like. The permit waits for all of them.

Who reviews what#

Reviewer	Looking at	Typical comment
Building and structural	Framing, lateral system, foundations, calculations	Load path not shown continuously to foundation
Energy	Compliance forms against the drawn assemblies	Glazing schedule does not match the compliance report
Plumbing, mechanical, electrical	Fixture counts, venting, equipment, service capacity	Panel schedule missing or load calculation absent
Zoning and planning	Setbacks, height, floor area, lot coverage, parking	Height measured from the wrong reference point
Fire	Access, hydrants, sprinklers, separations, egress	Turnaround dimensions not demonstrated
Grading and public works	Drainage, earthwork, right of way, encroachments	Drainage not shown discharging to an approved point
Accessibility	Applicable path of travel and clearances	Required clearances not dimensioned

Two features of that table cause most of the confusion. First, the reviewers do not talk to each other much, so a change made to satisfy one can create a comment from another. Second, several of them are checking your drawings against another document you submitted, not just against the code. The energy report, the structural calculations and the civil plans all have to agree with the architectural sheets, and the most common corrections in the whole process are simply internal inconsistencies between those sets.

Why the first submittal never comes back clean#

Assume it will not. Planning on anything else is how schedules get built wrong.

Completeness is judged, not measured. There is no checklist that guarantees approval. Reviewers have discretion about what constitutes a demonstrated condition, and different reviewers exercise it differently.

Consultants work on different clocks. The structural calculations, the energy report and the civil drawings are produced by separate offices at separate times. Unless someone reconciles them sheet by sheet before submittal, they will disagree somewhere, and disagreement is an automatic comment.

Codes are amended locally. State building code is adopted with local amendments, and those amendments are where an otherwise experienced design team gets caught. A detail that cleared review in one city may draw a comment in the next one over.

Corrections generate corrections. Revising a sheet to answer a comment changes a dimension that a different discipline was relying on. The second review then finds items the first review never reached, because the first review stopped at the more fundamental problem.

None of this is a sign that something has gone wrong. It is the ordinary operation of the process, and a first submittal that returns with a page of comments is a normal first submittal. Where this review sits among the other approvals a project needs is covered in the permitting topic.

Anatomy of a correction list#

A correction list is a numbered document, organized by discipline, where each item cites what is required and what the reviewer could not find. In practice the items fall into a few recognizable categories.

• Missing information. A detail, a schedule, a calculation or a note that the code requires and the set does not contain. The cheapest kind to fix.
• Inconsistency. Two sheets or two documents that contradict each other. Cheap to fix and embarrassing to receive, because it is self inflicted.
• Clarification. The reviewer cannot follow what is drawn and wants it shown differently. Often resolved with a single added detail.
• Substantive noncompliance. Something in the design does not meet a requirement: height, setback, egress, structural capacity. This is the category that can force a redesign.
• Local amendment or policy. A jurisdiction specific requirement that the design team did not apply.

The response is not just revised drawings. It is a written response letter that answers every numbered item, states what changed and points to the sheet where the change appears. Reviewers work from that letter. A resubmittal without one, or with a letter that answers six of nine comments, tends to come back with the same three comments and another full cycle attached to them.

Corrections versus a genuine code disagreement#

Most items on a list are not arguable. Something is missing, and you add it. Fighting a correction that could have been answered in an afternoon is the most expensive habit in permitting, because the disagreement costs a full cycle whether you win it or not.

A real disagreement is different, and it looks like one of these:

A factual reading dispute. The reviewer has applied a requirement to a condition it does not govern, or measured something from the wrong reference. These are usually resolved in a conversation, with the code section and the drawing open at the same time.

An alternate means request. The code permits approaches other than the prescriptive one when equivalent performance can be demonstrated. That is a formal submission with engineering behind it, not an argument, and it needs to be raised early because it takes its own review time.

A true appeal. Jurisdictions have a formal path to escalate a determination to a supervisor or a board. It exists, it is occasionally the right move, and it is slow. Reserve it for items that would otherwise force a redesign.

The judgment call is about cost. Compare what compliance would cost against a cycle of delay plus the effort of arguing. Drawing the detail is almost always cheaper than winning the point. The exception is a comment that would force a real change to the building, and that one is worth resolving properly and early, in a conversation with a proposal attached, rather than by resubmitting the same sheet with a note explaining why the reviewer is wrong.

The one lever that actually moves the calendar#

Everything inside a cycle is largely fixed. Queue time belongs to the department. Review time belongs to the reviewer. What belongs to you is how many cycles happen, and how fast you turn each one around.

A complete first submittal does not get reviewed faster. It gets reviewed fewer times, and since each avoided loop removes queue time, review time and response time together, that is where the months come from.

Practically, a complete submittal means: consultant documents reconciled against the architectural set before filing, local amendments researched for that specific jurisdiction, deferred submittals identified and labeled rather than left ambiguous, and someone who has read a correction list before doing a pre-submittal review of the whole package. On multi-department projects that last pass is handled as specialty services rather than left to whoever files the application.

The second lever is response speed. A correction list that sits for three weeks on a consultant's desk has added three weeks to the permit, and it is the one delay in the entire process nobody but you controls.

What this means for how you run a submittal#

Treat plan check as a documentation exercise with a scoreboard, and it becomes predictable. Assume corrections. Budget the cycles into the schedule rather than hoping to avoid them. Assign one person to own the correction list and the response letter so nothing waits on a consultant nobody called.

Then look at the drawings before they go in. The difference between a permit that issues in three months and one that issues in eight is rarely complexity. It is whether the set that entered the queue was actually finished, or whether the queue got used as a review service. A department will review whatever you hand it, as many times as you hand it over, and it will charge you a fresh place in line every time.