The path from a PDF drawing set to a submitted bill of quantities has more decision points than most estimating teams document. It is a process that experienced estimators carry in their heads, accumulated over years of practice, and that junior team members absorb by proximity. When that institutional knowledge is held informally, the workflow varies from estimator to estimator and from bid to bid.
What follows is the sequence that works for a general contractor on a mid-size commercial project. It is not a universal standard; it reflects what we have found produces consistent, auditable BOQs with a minimum of rework. We write it out explicitly because making the workflow visible is the first step toward making it faster.
Stage 1: Drawing set intake and completeness check
The first task when a tender package arrives is to verify that the drawing set is complete enough to take off. This is not a formality. Incomplete drawing sets are extremely common, and beginning a takeoff on an incomplete set means either restarting elements of the count after missing sheets arrive, or submitting a bid with unexplained exclusions that weaken the tender.
A basic completeness check covers: floor plans at each level (including basement and roof), at least one elevation per building face, representative sections through the primary structural system, structural drawings including floor and roof framing plans, a door schedule and window schedule, a finishes schedule, and MEP schematic drawings to the level needed for provisional sums or subcontract scope definition.
The completeness check also confirms drawing revisions. If the architectural and structural drawings are not at the same revision, there is a scope discrepancy risk. Structural drawings at an earlier revision than architectural drawings may not account for architectural changes made after the structural coordination stage. Document this before counting.
Estimated time for a 100-sheet set: 45 to 90 minutes. This time is often not counted as part of the takeoff schedule, but it should be. It is the step that determines whether the rest of the work can proceed on a clean basis.
Stage 2: BOQ structure definition
Before counting anything, define the output structure. A BOQ structured incorrectly at the outset produces counting work that cannot be assembled into the required format without reconstruction. The output structure should match the contract work sections used by the employer, the sub-trade breakdown needed for procurement, and the level of detail that can actually be extracted from the drawing set provided.
Standard BOQ structures for commercial construction in Korea follow work-section logic: substructure, frame, upper floors, roof, external walls, windows, internal walls, internal finishes, and services in provisional sums. Some tenders specify a different structure, often matching the employer's cost plan categories. Use the employer's structure if specified. Do not substitute your firm's standard template without checking that the categories align.
The structure definition step also includes identifying which quantities will be measured directly (concrete volumes, partition linear metres, floor areas) versus which will be covered by lump sums or provisional sums pending subcontractor quotes (specialist mechanical, electrical, lifts, facade engineering). Drawing the boundary between directly measured and provisional items affects what you need to count and what you do not.
Stage 3: Primary measurement pass
With the BOQ structure defined, the counting begins. The most efficient sequence moves through the building from the ground up and from the structural shell inward to finishes:
Substructure and foundations: Count from structural drawings. Ground floor slab area, pile quantities if shown, foundation types and locations. Foundation drawings often lag the rest of the set, so record what is available and note any assumptions.
Structural frame: Columns by cross-section type and floor level from structural plans and column schedules. Beams by span and cross-section category. Structural walls (shear walls and core walls) separately from non-structural walls, as they have different rates and often different measurement rules.
Floor slabs and roof: Area by level from structural drawings or architectural floor plans where the structural plan is not disaggregated by level. Note any slab thickness changes and voided areas over 1 square metre, which are typically deducted.
External envelope: Cladding area from elevations, window area from elevations and window schedules, external door quantities. This is where the cross-check between elevation measurements and plan perimeter measurements matters most. Use elevations as the primary source.
Internal partitions: Linear metres by partition type from floor plans. This step requires cross-referencing the fire strategy drawings to flag fire-rated sections (see the separate article on estimation mistakes). Partition heights require section drawings or a schedule.
Finishes: Floor area by finish type from floor plans and finishes schedules, wall finish area from room schedules if provided (less common in commercial tenders), ceiling area and type by zone.
Fixtures and fittings: Count from floor plans and specification schedules, cross-checked between sources as discussed.
Stage 4: Cross-referencing and gap resolution
After the primary measurement pass, spend a focused session on cross-referencing. Check counts made from schedules against counts from plans. Check measurements from plans against measurements from sections. Record every discrepancy and resolve each one before closing the count.
Gap resolution covers: elements that appear in the specification but cannot be located in the drawings (request a drawing query list to the design team); elements detected on drawings that are not in any schedule (flag as unclassified items with a provisional rate); scope items called for in the preliminaries or specification preambles that have no corresponding drawing element (often refers to site setup, protection works, and temporary structures).
This stage is where the senior estimator's judgment adds the most value. A junior team member can follow the measurement sequence above. The cross-referencing decisions, specifically which discrepancy reflects a design error and which reflects a legitimate scope ambiguity requiring a tender query, require construction knowledge that cannot be fully systematised.
Stage 5: Rate application and BOQ assembly
Measured quantities go into the BOQ structure with preliminary rates applied. At this stage the BOQ is a quantity document, not a priced document. The quantity is locked; the rates are reviewed separately and may be revised based on current material and labour market conditions.
BOQ assembly involves: aggregating floor-level counts to project totals where the BOQ requires project-level quantities; splitting quantities where the BOQ structure has sub-categories that require further breakdown from the primary count; applying wastage factors to material quantities (these should be applied at rate stage, not at quantity stage, to keep the measured quantity auditable).
The assembled BOQ goes through a brief internal review before pricing. The review checks for internal consistency: are the partition quantities consistent with the floor areas? Are the fixture counts plausible for the building type? Is any work section missing that should be present based on the project description? This review catches errors that can be confirmed without revisiting the drawings.
Where automation fits in this workflow
Stages 1 and 2 are judgment-intensive. Completeness checking requires knowing what should be present. BOQ structure definition requires knowing how the output will be used. These are not candidates for automation in their current form.
Stage 3, the primary measurement pass, is where the largest block of time is consumed by routine pattern recognition: finding the door symbol, measuring the wall line, tracing the slab boundary. These are tasks where the drawing layer and symbol are consistent, the scale is defined, and the count follows a repeatable rule. This is the stage where systems like FOBECON reduce the time cost most significantly.
Stage 4 is partially automatable. Cross-checking a count made from a floor plan against a count made from a schedule is a comparison that a system can perform faster than a person. The judgment about which discrepancy is a design error versus a scope ambiguity is not automatable. The system surfaces the discrepancy; the estimator resolves it.
Stage 5 is largely a data assembly task and benefits from having Stage 3 outputs in a structured format that can be mapped directly to BOQ line items without re-entry.
The workflow above is not new. Experienced estimating teams have followed some version of it for decades. What changes when Stage 3 is handled by a system is the time distribution across stages: the hours that went to symbol counting are reallocated to cross-referencing and scope judgment, which is where the senior estimator's value actually lies.