What "From Drawing / Scheduled / Counted / Derived / Estimated" Mean on a BOM
When a takeoff lands on your desk, the pressing question isn't "what's the total" — it's "which numbers can I use as-is, and which do I need to look at myself?"
Two numbers in the same spreadsheet can be worlds apart in reliability. A section read straight off a schedule and a section the estimator assumed look identical in a cell — but one is almost never wrong and the other can be a full size off.
So every row carries two bases.
Two axes, kept separate
Section basis — how the section was determined:
| Value | Meaning |
|---|---|
| From drawing | Stated on a schedule or the drawing face |
| Estimated | Not stated; assumed from span, load or similar members |
Quantity basis — how the count was determined:
| Value | Meaning | Typical error |
|---|---|---|
| Scheduled | Count given in a schedule | ~0 |
| Counted | Counted off the plan | 1–3% |
| Derived | From spacing and bay count | 5–15% |
| Estimated | No basis on the drawings | 15%+ |
These are independent. A very common case: the section is from the drawing (the schedule gives C180×70×20×2.5) but the count is derived (the drawing only gives @1200 spacing). Trust the section; check the count.
Why not collapse this into one confidence score
A single percentage reads cleaner but can't be acted on. A reviewer needs to know what to go look at:
- Section in doubt → check the schedule, check the details
- Count in doubt → check the plan, count it again
Collapse them and that's gone. You're left with "85% confident" and no idea which sheet to open.
How to review
- Look at the review tonnage as a share of total. Normally 2–8%. Above 15% means
the drawings themselves are incomplete, or too much was assumed.
- Only review the derived and estimated rows. From-drawing and scheduled rows carry
most of the tonnage but almost never move.
- Sort by weight. A 49-tonne derived row is worth twenty minutes. A 0.3-tonne one
is worth a glance.
- Cross-check. Column counts reconcile against anchor bolts; bay counts against the
grid. Two sources agreeing promotes "counted" to effectively reliable.
A real example
A plant project: 2011.5 tonnes across 142 rows.
| Basis pair | Tonnes | Share |
|---|---|---|
| From drawing + Scheduled | 1967.3 | 97.8% |
| From drawing + Derived | 27.2 | 1.4% |
| Estimated + Counted | 17.0 | 0.8% |
44.2 tonnes to review — 2.2%. The reviewer opens 3 rows, not 142.
Why this is worth being strict about
Takeoff numbers become purchase orders and bids. A 5% error on a multi-million contract is real money — and you don't know which row it's in unless every row states where it came from.
Stating it is cheap. Not stating it is expensive.