The area is the easy part of a takeoff
Practitioner reference · steep-slope asphalt worked in full, principles transfer
Two roofs with the same footprint, the same pitch and the same 13.42 squares. One has no angled cuts and 40 linear feet of cap. The other has 108 feet of hip and valley and 110 feet of cap. A flat waste percentage prices neither of them correctly.
Why do some jobs land on material and others quietly bleed bundles?
Your takeoff has three outputs, not one: surface area, linear footage by edge type, and counts. Area is the easy part and the part software gets right. The money leaks out of the other two — hip, valley, ridge, starter, drip edge, boots, fasteners — and out of a flat waste percentage applied to a roof whose geometry does not deserve it.
Same squares, different jobSection link
Two roofs enclosing 1,200 sq ft of plan area, every plane at 6:12. Every figure below is derived from that geometry, not surveyed. Roof A is a 40 ft × 30 ft gable; Roof B is a hipped 40 ft × 24 ft main wing with a 24 ft × 10 ft cross gable.
| Quantity | Roof A — simple gable | Roof B — hip and cross gable | Difference |
|---|---|---|---|
| Plan area | 1,200 sq ft | 1,200 sq ft | none |
| Surface area | 1,341.6 sq ft (13.42 sq) | 1,341.6 sq ft (13.42 sq) | none |
| Eave | 80.0 ft | 124.0 ft | +44 ft |
| Rake | 67.1 ft | 26.8 ft | 40.2 ft less on B |
| Ridge | 40.0 ft | 38.0 ft | 2.0 ft less on B |
| Hip | 0 ft | 72.0 ft | +72.0 ft |
| Valley | 0 ft | 36.0 ft | +36.0 ft |
| Hip and ridge cap | 40.0 ft | 110.0 ft | 2.8× as much |
| Drip edge and starter | 147.1 ft | 150.8 ft | +3.8 ft |
| Modelled offcut waste, no reuse | 8.5% | 23.8% | +15.3 points on B |
Read this table one item at a time
Plan area
- Roof A — simple gable
- 1,200 sq ft
- Roof B — hip and cross gable
- 1,200 sq ft
- Difference
- none
Surface area
- Roof A — simple gable
- 1,341.6 sq ft (13.42 sq)
- Roof B — hip and cross gable
- 1,341.6 sq ft (13.42 sq)
- Difference
- none
Eave
- Roof A — simple gable
- 80.0 ft
- Roof B — hip and cross gable
- 124.0 ft
- Difference
- +44 ft
Rake
- Roof A — simple gable
- 67.1 ft
- Roof B — hip and cross gable
- 26.8 ft
- Difference
- 40.2 ft less on B
Ridge
- Roof A — simple gable
- 40.0 ft
- Roof B — hip and cross gable
- 38.0 ft
- Difference
- 2.0 ft less on B
Hip
- Roof A — simple gable
- 0 ft
- Roof B — hip and cross gable
- 72.0 ft
- Difference
- +72.0 ft
Valley
- Roof A — simple gable
- 0 ft
- Roof B — hip and cross gable
- 36.0 ft
- Difference
- +36.0 ft
Hip and ridge cap
- Roof A — simple gable
- 40.0 ft
- Roof B — hip and cross gable
- 110.0 ft
- Difference
- 2.8× as much
Drip edge and starter
- Roof A — simple gable
- 147.1 ft
- Roof B — hip and cross gable
- 150.8 ft
- Difference
- +3.8 ft
Modelled offcut waste, no reuse
- Roof A — simple gable
- 8.5%
- Roof B — hip and cross gable
- 23.8%
- Difference
- +15.3 points on B
Surface area is plan area × the pitch factor for 6:12. Hip and valley lengths use the hip multiplier on the common run, which is exactly 1.500 at 6:12. Modelled offcut waste is the zero-reuse upper bound explained further down the page; it is a model, not a measurement.
Where this page is useful, and where it is notSection link
This page is a model and a checklist, not a system of record. Here is where it earns its keep and where it does not.
Best when
- You already have a repeatable method and want to find the leak in it, rather than learning a first one.
- You bid enough cut-up roofs that a systematic waste error compounds across a season.
- You quote in squares and want the measured area and the ordered area to stop being the same number on your proposal.
- You want a defensible way to explain to a customer why two roofs of the same size are not the same price.
Think twice if
- You already job-cost by phase and know your own returned-and-leftover material by roof type. Your own history beats every model on this page, including the one it spends the most space on.
- You work a single repeatable product — tract housing, one shape, one pitch — where a calibrated flat percentage is genuinely correct and cheaper to run.
- You are a low-slope-only contractor. The geometry here is steep-slope. Membrane takeoff turns on seam length, insulation layout, fastener density and tapered design, which live under commercial roofing.
- You are a homeowner. This page assumes you can install a roof, and it will not tell you whether yours needs one.
What changes the answer
- Whether your crews are paid by the square. Piece rate moves the incentive on offcut reuse, and the model assumes an average crew rather than yours.
- Whether the covering can be reordered mid-job. A discontinued or long-lead colour makes deliberate over-ordering the cheap decision, and the model does not know that.
- Whether the roof is a tear-off or new construction. Tear-off adds layer count, deck condition, disposal weight and a discovery risk that no measurement source can see.
- Your supplier's return policy. Full-bundle returns at a restocking fee change the arithmetic of rounding up; no returns changes it in the other direction.
A takeoff is three numbers, and only one of them is areaSection link
Area, linear footage, and counts. Three outputs, three different ways to be wrong about them.
Every takeoff produces three families of number, and they fail differently.
Area is the one everyone checks. It is also the one that is now close to solved: a plan set, a traced aerial outline and a careful field measurement of the same roof usually land close enough together that the gap does not decide the bid, and when they do not, the cause is almost never precision. This page publishes no figure for how closely they agree, because the only accuracy number it could read is one vendor’s, about its own product, and it is quoted and qualified further down. The taxonomy of why two measurements of one roof disagree — scope, waste already folded in, overhang, slope assignment, imagery date, obstruction, projected versus along-the-slope — is worked through on how a roof gets measured and is not repeated here.
Linear footage by edge type is where estimates start to diverge. Eave, rake, ridge, hip and valley are five different products, five different labour rates and, in the case of hip and valley, the driver of how much material ends up in the dumpster. Software will hand you these lengths. Whether your estimate consumes all five of them is a different question.
Counts are the ones that vanish. Pipe boots, step-flashing pieces, ridge vent sections, exhaust terminations, satellite mounts, skylights, chimney crickets, and the fasteners underneath all of it. Individually each is a rounding error. Collectively, on a cut-up roof, they are the difference between a job that hits its material budget and one that does not.
The two roofs this page argues with
Both enclose 1,200 square feet of plan area. Both are 6:12 throughout. Multiply plan area by the pitch factor — √(1 + (6 ÷ 12)²), which is 1.118, and which is derived on roof pitch rather than here — and both carry 1,341.6 square feet of surface area. That is 13.42 squares each.
They are not the same job, and nothing about the area number says so.
Three ways to get a measurement, and what each one structurally cannot seeSection link
Not which is most accurate. Which errors each one produces, and which of those errors cost money.
Accuracy is the wrong axis. All three sources are accurate enough about geometry for a bid. They differ in what they are blind to, and the blind spots are where the losses are.
| Source | What it is genuinely good at | Its characteristic error | What it cannot see at all |
|---|---|---|---|
| Plans and drawings | Plane dimensions, slopes, and the design intent behind them. On new construction it is the only source that exists before the roof does. | Design intent is not as-built. Framing gets changed, overhangs get adjusted in the field, and the plane count on the roof plan and on the elevations do not always agree with each other. | Anything about an existing roof: layers, deck condition, prior repairs, or what a previous crew left behind. |
| Aerial and satellite reports | Fast, repeatable, plane-by-plane geometry with a linear schedule, produced without anyone leaving the ground. On complex geometry it is better than most people are with a tape. | It measures the top surface as the imagery saw it. Tree canopy and deep shadow move edges. A plane framed between two whole-inch slopes gets assigned to one of them. Reported total squares sometimes already includes a waste allowance and sometimes does not. | Layer count. Deck condition. Ventilation as built. Access, staging and disposal. And anything built, removed or re-covered since the capture date. |
| Field measurement | The only source that sees the job rather than the roof: layers at the rake edge, soft deck underfoot, existing flashing condition, penetration count, and where a container can actually sit. | Human transcription. A plane missed, a dimension written down wrong, or an overhang measured to the wall on one plane and to the eave on another. Slower, and on cut-up roofs less reliable than software for pure geometry. | Nothing about the visible roof — but it costs roof access, which is the exposure the federal fall-protection standard exists to address. The safety notice at the top of this page is not decorative. |
Read this table one item at a time
Plans and drawings
- What it is genuinely good at
- Plane dimensions, slopes, and the design intent behind them. On new construction it is the only source that exists before the roof does.
- Its characteristic error
- Design intent is not as-built. Framing gets changed, overhangs get adjusted in the field, and the plane count on the roof plan and on the elevations do not always agree with each other.
- What it cannot see at all
- Anything about an existing roof: layers, deck condition, prior repairs, or what a previous crew left behind.
Aerial and satellite reports
- What it is genuinely good at
- Fast, repeatable, plane-by-plane geometry with a linear schedule, produced without anyone leaving the ground. On complex geometry it is better than most people are with a tape.
- Its characteristic error
- It measures the top surface as the imagery saw it. Tree canopy and deep shadow move edges. A plane framed between two whole-inch slopes gets assigned to one of them. Reported total squares sometimes already includes a waste allowance and sometimes does not.
- What it cannot see at all
- Layer count. Deck condition. Ventilation as built. Access, staging and disposal. And anything built, removed or re-covered since the capture date.
Field measurement
- What it is genuinely good at
- The only source that sees the job rather than the roof: layers at the rake edge, soft deck underfoot, existing flashing condition, penetration count, and where a container can actually sit.
- Its characteristic error
- Human transcription. A plane missed, a dimension written down wrong, or an overhang measured to the wall on one plane and to the eave on another. Slower, and on cut-up roofs less reliable than software for pure geometry.
- What it cannot see at all
- Nothing about the visible roof — but it costs roof access, which is the exposure the federal fall-protection standard exists to address. The safety notice at the top of this page is not decorative.
A measurement source produces a roof. A takeoff needs a job. Everything in the last column has to be collected some other way, and the fact that two of these sources can be run without visiting the property is exactly why that column gets skipped.
The productive combination for a re-roof is usually not one of these. It is a commissioned or traced report for the geometry, plus a ground and eave-level visit for everything in the fourth column, plus a look at the condition of what is there. The report is doing the arithmetic; the visit is doing the estimating.
Waste is a function of cut length, not of areaSection link
A flat percentage overprices a simple roof and underprices a cut-up one. Here is the derivation, and here is what it does to the two roofs above.
Offcut waste comes from one place: a rectangle being cut to meet a line that is not parallel to it. Everything else people call waste — breakage, damaged bundles, a colour that had to be replaced, the bundle left in the garage — is real but is not a function of the geometry, and lumping it in with offcuts is what makes a single percentage look defensible.
The derivation
Take a cut line on a roof plane — a rake, a hip, a valley. Courses run horizontally into it, and each course has to be trimmed. Over successive courses the joint pattern advances against the cut line, so the length of shingle left in the installer’s hand is effectively uniform across the shingle length and averages half of it.
The number of courses a cut crosses is its in-plane rise divided by the exposure. Each of those courses loses, on average, half a shingle’s length of coverage — which is (L ÷ 2) × exposure of area. So:
lost coverage = (in-plane rise ÷ exposure) × (L ÷ 2) × exposure = in-plane rise × L ÷ 2
Exposure cancels. And because in-plane rise is the plan up-slope extent times the pitch factor, while surface area is plan area times the same pitch factor, the pitch cancels out of the percentage as well:
offcut waste fraction = (Σ plan up-slope extent of every cut × L ÷ 2) ÷ plan area
That is the whole model. Two assumptions carry it: that the leftover averages half a shingle, and that the leftover is not reused. The first is a property of the geometry. The second is a property of the crew, and it is the one worth arguing about — which is why the numbers below are given as a range across reuse rates rather than as a single figure.
The one thing that decides whether an offcut is reusable
At a rake, the cut is straight up the slope and the offcut is a full-height rectangle. It is exactly what starts the next course at the far rake. Reuse is high, and on a plain gable it approaches complete.
At a hip or a valley, the cut is diagonal and the offcut has a bevelled end. It fits a diagonal of the same hand somewhere else on the roof, or it fits nowhere. Reuse is low, it takes deliberate stocking to achieve, and it is the first discipline to go when the crew is behind.
That is the entire mechanism behind “cut-up roofs waste more.” It is not that there is more cutting per square foot in some vague sense. It is that the offcuts are the wrong shape.
One thing the geometry does not decide for you: how many sides of a valley get cut is a function of the valley method. An open valley is cut on both planes. A closed-cut valley is cut on one, because the shingles from the other plane run through it. A woven valley is cut on neither. Everything below assumes both sides on both of Roof B’s valleys, which is the open-valley case and the upper bound. That assumption is worth 6.8 points of Roof B’s modelled waste, so halve the valley contribution before you use the number if you run closed-cut.
The two roofs, run through it
Using IKO Cambridge dimensions — 40.875 in long, 5.875 in exposure — as the declared input, and reading the plan up-slope extent of every cut line off the geometry above. Roof A has four rakes, each with 15 ft of plan up-slope extent. Roof B has two rakes at 12 ft each, plus four hips and two valleys, each of which is cut on both sides with 12 ft of plan up-slope extent per side.
| Offcut reuse assumption | Roof A — simple gable | Roof B — hip and cross gable | What the gap means |
|---|---|---|---|
| Nothing reused | 8.5% | 23.8% | The upper bound. Nothing is reused; every trimmed piece is dropped. |
| Half of every offcut reused | 4.3% | 11.9% | Half of everything reused, regardless of shape. Flattering to the cut-up roof and unrealistic in its favour. |
| Three quarters of rake offcuts, a quarter of angled ones | 2.1% | 16.2% | The realistic shape of it: rectangular rake offcuts mostly used, bevelled hip and valley offcuts mostly not. |
Read this table one item at a time
Nothing reused
- Roof A — simple gable
- 8.5%
- Roof B — hip and cross gable
- 23.8%
- What the gap means
- The upper bound. Nothing is reused; every trimmed piece is dropped.
Half of every offcut reused
- Roof A — simple gable
- 4.3%
- Roof B — hip and cross gable
- 11.9%
- What the gap means
- Half of everything reused, regardless of shape. Flattering to the cut-up roof and unrealistic in its favour.
Three quarters of rake offcuts, a quarter of angled ones
- Roof A — simple gable
- 2.1%
- Roof B — hip and cross gable
- 16.2%
- What the gap means
- The realistic shape of it: rectangular rake offcuts mostly used, bevelled hip and valley offcuts mostly not.
All figures are modelled offcut waste only. Breakage, damaged bundles, colour replacement, and the bundle left behind for the owner are additional and are not a function of geometry. Roof A's waste is entirely rake waste (8.5% at zero reuse). Roof B's splits into 3.4% rake and 20.4% hip and valley — of which 6.8 points is valley, assumed cut on both planes, which is the open-valley case.
Read the bottom row. A flat ten per cent applied to both roofs overprices Roof A by roughly eight points of material and underprices Roof B by roughly six. The overpricing is invisible — it loses bids you never hear about. The underpricing is invisible too, until job costing shows material over on complex work and the company decides its crews are wasteful.
The useful move is not a better single number. It is to stop deriving waste from area at all and derive it from the linear schedule the takeoff already produced: total the plan up-slope extent of every cut, separate the straight cuts from the angled ones, and apply your own reuse assumption to each. It is two more cells in a spreadsheet.
Where this model is wrong
- On very small or very simple roofs it understates. When offcut waste is two per cent, a single damaged bundle and the rounding to whole bundles both matter more than the geometry does.
- It assumes the leftover distribution is even. A strict joint offset against a particular cut angle can make the leftover length nearly constant course after course — much better than half a shingle if the constant is small, much worse if it is not.
- It assumes every angled line is cut on both sides. True of a hip and of an open valley. A closed-cut valley is cut on one plane and a woven valley on neither, so the valley method — a workmanship decision, not a measurement — moves Roof B’s modelled waste by about 3.4 points on its own.
- It says nothing about labour, and labour is the larger number. An angled cut costs a knife stroke and a positioning move as well as material. The productivity effect of complex geometry is real and is not modelled here.
- It is not a substitute for your own history. If you record returned and leftover material by roof type, that data outranks this model completely. Use it.
Accessories are individually trivial and collectively decisiveSection link
Each of these looks like a rounding error. Together, on Roof B, they are a meaningful share of the material order and most of the detail labour.
The pattern is consistent: quantities derived from area get estimated carefully, and quantities derived from linear footage or from counts get estimated from habit. Habit was formed on the last roof, and the last roof was a different shape.
| Quantity | What it is actually derived from | Why it goes missing | On Roof A / Roof B |
|---|---|---|---|
| Starter at the rake | Eave linear footage plus rake linear footage. The Department of Energy's guide puts starter at both, lapping three inches where they meet. | It is ordered off eave length because starter is thought of as an eave product. | 147.1 ft / 150.8 ft — and on Roof A the rake alone is 67.1 ft of it |
| Drip edge at eaves and rakes | The same eave-plus-rake total as starter, plus corner and splice allowance. | Ordered in tens of feet from a mental image of the front elevation rather than from the schedule. | 147.1 ft / 150.8 ft |
| Hip and ridge cap | Ridge length plus hip length, divided by the published coverage of the specific cap product. | Estimated from ridge alone, or divided by a remembered coverage from a different product. | 40.0 ft / 110.0 ft — 2 bundles against 4 at 36.5 ft per bundle, or 2 against 6 at 20 |
| Valley lining | Valley length, and whether the valley is open or closed. Open valleys need lining metal; closed valleys need lining membrane and a cut discipline. | There is no valley on the last three roofs, so there is no line for it on the estimate template. | 0 ft / 36.0 ft |
| Ice barrier courses | Eave length × the up-slope reach to a point 24 in inside the interior plane of the exterior wall, which is how the Department of Energy's guide puts it, then divided into courses by the roll width net of its side lap. | One course is assumed because one roll width sounds like enough. At 6:12 with any real overhang it is not. | 80 ft of eave / 124 ft — 55% more on the roof with the same area |
| Step flashing pieces | Sidewall run length converted up the slope and divided by the exposure — one piece per course, not one piece per foot. | Estimated as linear feet of wall and then bought in feet rather than in pieces. | 20 ft of sidewall measured in plan, at 6:12, is 46 pieces — not 20 |
| Penetration boots and terminations | A count off the roof plane photograph: plumbing vents, bath fans terminating through the roof, range and dryer terminations, radon stacks, abandoned stubs, satellite and antenna mounts. | Each is individually cheap, so each is individually skipped, and nobody adds them up. | Not a function of geometry — count them or absorb them |
| Fasteners | Shingle count × the fastener count the product and the wind requirement call for, plus cap nails for underlayment, drip edge fasteners, and cap shingle fasteners. | Treated as a consumable rather than a quantity, then bought by the box on feel. | 805 shingles either way; four nails each is 3,220, six is 4,830 |
| Ridge vent, separately from cap | Ridge length again, as a separate quantity from cap: the vent occupies the ridge and the cap goes over the vent. | Counted once, as if vent and cap were the same line. | 40.0 ft / 38.0 ft of ridge, plus the hips that carry cap but no vent |
| Tear-off weight and container count | Layer count × area × the weight per square of what is coming off, then converted to containers and to overweight risk. | Estimated from area alone, because layer count was never established. | Not visible in any measurement source — check the rake edge |
| Access, staging and protection | The site: distance from drop to ladder, storeys, jack courses on steep planes, sheets of ground protection, permit for a container on the street. | It is not on the roof, so it is not in the roof takeoff. | Identical roofs on different lots are different jobs |
| Deck replacement | A unit price and an included quantity, agreed in writing before tear-off. | Written as an open allowance, or not written at all, and then negotiated on day two with the deck open. | Set the unit price; do not set the quantity from a guess |
Read this table one item at a time
Starter at the rake
- What it is actually derived from
- Eave linear footage plus rake linear footage. The Department of Energy's guide puts starter at both, lapping three inches where they meet.
- Why it goes missing
- It is ordered off eave length because starter is thought of as an eave product.
- On Roof A / Roof B
- 147.1 ft / 150.8 ft — and on Roof A the rake alone is 67.1 ft of it
Drip edge at eaves and rakes
- What it is actually derived from
- The same eave-plus-rake total as starter, plus corner and splice allowance.
- Why it goes missing
- Ordered in tens of feet from a mental image of the front elevation rather than from the schedule.
- On Roof A / Roof B
- 147.1 ft / 150.8 ft
Hip and ridge cap
- What it is actually derived from
- Ridge length plus hip length, divided by the published coverage of the specific cap product.
- Why it goes missing
- Estimated from ridge alone, or divided by a remembered coverage from a different product.
- On Roof A / Roof B
- 40.0 ft / 110.0 ft — 2 bundles against 4 at 36.5 ft per bundle, or 2 against 6 at 20
Valley lining
- What it is actually derived from
- Valley length, and whether the valley is open or closed. Open valleys need lining metal; closed valleys need lining membrane and a cut discipline.
- Why it goes missing
- There is no valley on the last three roofs, so there is no line for it on the estimate template.
- On Roof A / Roof B
- 0 ft / 36.0 ft
Ice barrier courses
- What it is actually derived from
- Eave length × the up-slope reach to a point 24 in inside the interior plane of the exterior wall, which is how the Department of Energy's guide puts it, then divided into courses by the roll width net of its side lap.
- Why it goes missing
- One course is assumed because one roll width sounds like enough. At 6:12 with any real overhang it is not.
- On Roof A / Roof B
- 80 ft of eave / 124 ft — 55% more on the roof with the same area
Step flashing pieces
- What it is actually derived from
- Sidewall run length converted up the slope and divided by the exposure — one piece per course, not one piece per foot.
- Why it goes missing
- Estimated as linear feet of wall and then bought in feet rather than in pieces.
- On Roof A / Roof B
- 20 ft of sidewall measured in plan, at 6:12, is 46 pieces — not 20
Penetration boots and terminations
- What it is actually derived from
- A count off the roof plane photograph: plumbing vents, bath fans terminating through the roof, range and dryer terminations, radon stacks, abandoned stubs, satellite and antenna mounts.
- Why it goes missing
- Each is individually cheap, so each is individually skipped, and nobody adds them up.
- On Roof A / Roof B
- Not a function of geometry — count them or absorb them
Fasteners
- What it is actually derived from
- Shingle count × the fastener count the product and the wind requirement call for, plus cap nails for underlayment, drip edge fasteners, and cap shingle fasteners.
- Why it goes missing
- Treated as a consumable rather than a quantity, then bought by the box on feel.
- On Roof A / Roof B
- 805 shingles either way; four nails each is 3,220, six is 4,830
Ridge vent, separately from cap
- What it is actually derived from
- Ridge length again, as a separate quantity from cap: the vent occupies the ridge and the cap goes over the vent.
- Why it goes missing
- Counted once, as if vent and cap were the same line.
- On Roof A / Roof B
- 40.0 ft / 38.0 ft of ridge, plus the hips that carry cap but no vent
Tear-off weight and container count
- What it is actually derived from
- Layer count × area × the weight per square of what is coming off, then converted to containers and to overweight risk.
- Why it goes missing
- Estimated from area alone, because layer count was never established.
- On Roof A / Roof B
- Not visible in any measurement source — check the rake edge
Access, staging and protection
- What it is actually derived from
- The site: distance from drop to ladder, storeys, jack courses on steep planes, sheets of ground protection, permit for a container on the street.
- Why it goes missing
- It is not on the roof, so it is not in the roof takeoff.
- On Roof A / Roof B
- Identical roofs on different lots are different jobs
Deck replacement
- What it is actually derived from
- A unit price and an included quantity, agreed in writing before tear-off.
- Why it goes missing
- Written as an open allowance, or not written at all, and then negotiated on day two with the deck open.
- On Roof A / Roof B
- Set the unit price; do not set the quantity from a guess
Cap bundle counts use IKO's published coverage for two of its own products: 36.5 linear feet per bundle for Hip & Ridge 12 and 20 linear feet per bundle for UltraHP. They are illustrative of the spread, not a specification for any other product.
Measured surface area is not order area, and both belong on the proposalSection link
They differ by waste and then again by the fact that material is sold in whole bundles.
Surface area is what the roof is. Order area is what gets bought. The conversion runs through the waste factor and then through the fact that a supplier sells whole bundles; the unit arithmetic itself is on roofing squares and is not repeated here.
Run the two roofs at the rates the bottom row of the offcut-waste table above suggests, rounded to whole percentages — five per cent for Roof A against a modelled 2.1 per cent, which leaves the simple roof room for breakage and a damaged bundle, and sixteen for Roof B against a modelled 16.2 — against a product covering 33.3 square feet per bundle:
- Roof A. 1,341.6 sq ft × 1.05 = 1,408.7 sq ft, which is 42.30 bundles, so 43 bundles, so 1,431.9 sq ft of order area. That is 6.7 per cent above measured.
- Roof B. 1,341.6 sq ft × 1.16 = 1,556.3 sq ft, which is 46.74 bundles, so 47 bundles, so 1,565.1 sq ft of order area. That is 16.7 per cent above measured.
The rounding is not nothing, and it hurts more the smaller the roof. One bundle is 33.3 square feet whatever the roof is: on these 13.42 squares that is 2.5 per cent, and on a six-square garage it is 5.5 per cent — larger than the whole offcut allowance a simple garage roof deserves. That is a good argument for pricing small jobs off bundles rather than off squares, and for tracking the leftover.
The same rounding runs through every accessory. Cap on Roof B is 3.01 bundles of a product covering 36.5 linear feet — which is four bundles, not three, and an estimate that took the quotient at face value is a bundle short on the roof at four in the afternoon.
Where your interests and the customer's genuinely divergeSection link
This site is a buyer's resource. It is not your business partner, and pretending the two sides always want the same thing would make this page useless to both.
Most of what is above serves everyone. An accurate takeoff means you do not eat a shortfall and the customer does not fund a padded one. But there are three places where the incentives actually pull apart, and they are worth naming rather than talking around.
The waste factor is an invisible price lever
Waste is a real cost and it belongs in the price. It is also the easiest number on a proposal to move without anyone noticing, because the customer sees a total in squares and cannot tell whether that figure is measured area or ordered area. This site tells buyers to ask which one it is, and to ask for both. That question is not an accusation; a contractor who can answer it in one sentence looks better than one who cannot, which is why stating both numbers is in your interest too.
The deck allowance can be a profit centre or a protection
A unit price for deck replacement, agreed before the tear-off, with a stated included quantity, protects both sides from the same discovery. A unit price set well above cost, applied to a quantity nobody can verify from the ground, is a different instrument. Buyers are told to ask for the unit price in writing and to ask how the quantity will be documented — photographs, sheet count, an on-site agreement before replacement. Estimators who already do that will find the question easy.
Leftover material belongs to whoever the contract says it does
The customer paid for the order area, waste and all. Whether the surplus stays on site is a contract term, not a convention, and it goes both ways: material left for future repairs is genuinely valuable to the owner, and hauling it away without saying so reads badly for a reason. Put it in the scope in one line and it stops being an argument.
What this page will not do, in any circumstance, is help anyone present an order quantity as a measurement, or a waste allowance as roof area. Those are the same sentence said two ways, and the page that teaches buyers to compare quotes already tells them to look for exactly that.
What changes these numbers on a real buildingSection link
The variables that move a takeoff, in the order they usually move it.
- Slope and drainage
Pitch multiplies area, ladder time, staging, and the labour rate per square. It does not, under the offcut model on this page, change the waste percentage at all: a steeper plane has proportionally more material and proportionally longer cuts, and the pitch factor cancels. What steepness genuinely adds is breakage, handling loss, roof jacks, and the time to move bundles — costs that belong in labour and equipment, not in the waste factor. Estimators who bump waste for steep pitch are usually pricing something real in the wrong line.
- Access and site conditions
Two-storey work, tight urban lots, no driveway, a dumpster that cannot get within a hundred feet of the ladder, a conveyor that has to be set twice, landscaping that has to be sheeted, and overhead service drops. None of these appear on any measurement of the roof, and all of them are quantities: ladder sets, jack courses, sheets of protection, hours of carry. Walk the ground before you price the roof.
- Code and jurisdiction
A cold-climate ice barrier is the quantity most often ordered one course short. The Department of Energy’s guidance describes a barrier running from the lowest edges of all roof surfaces to a point not less than 24 inches inside the interior plane of the exterior wall. With a 16-inch horizontal overhang and a wall assembly about 7.5 inches thick, that is 47.5 inches in plan, which at 6:12 is 53 inches up the slope. A 36-inch roll nets rather less than 36 inches once the side lap the product requires is taken out. That is two courses, not one, and it is two courses along 124 feet of eave on Roof B against 80 on Roof A.
That 24-inch figure is the Department of Energy's guidance for cold climates, not a code determination. Ice-barrier requirements depend on the edition adopted where the work is, on local amendments, on the climate designation, and on the existing assembly. This page publishes no code text and no jurisdiction's requirement; confirm the adopted edition, its amendments and its effective date with the authority having jurisdiction.- Wind
Fastener count is a takeoff quantity, not a supply-house afterthought. At 1.67 square feet of coverage per shingle — 40.875 in × 5.875 in — 1,341.6 square feet of roof is about 805 shingles. IKO publishes 33.3 square feet of coverage per bundle for that product and publishes no shingle count at all; twenty is simply that coverage divided by the shingle’s own dimensions, and it is a number to check against the bundle in front of you rather than to carry to another product. The Department of Energy’s asphalt-shingle guide says six nails are generally required to fasten one shingle strip, and six specifically in high-wind areas. Where a product’s baseline specification is four, moving to six is 1,610 additional fasteners on this roof and fifty per cent more nailing time — before any cap nails for underlayment, drip edge fasteners, or the cap shingles.
Fastener count, pattern, and placement within the nailing zone are set by the installation instructions for the specific product and by the wind requirement that applies to the site. A marketing wind rating is not a code determination, and this page does not publish a universal nailing method.- Low-slope work
Everything numeric on this page is worked in steep-slope asphalt because that is where the arithmetic is most transferable. On low-slope work the equivalent leaks are seam linear footage, insulation board count against a tapered layout, fastener and plate density by zone, termination and edge-metal footage, and what the existing assembly turns out to contain. Those live under roof condition assessment and recover versus tear-off.
Where a takeoff decision becomes a warranty questionSection link
Two places where a quantity decision becomes a coverage question.
- Component substitution is an estimating decision
Enhanced or system warranty programmes generally condition eligibility on the components used, and those components are named in the programme document rather than inferred. That makes accessory selection a takeoff decision and not a purchasing decision: substituting a cheaper starter or cap after the estimate is priced can change what the completed roof qualifies for. Whether it does, in a given programme, is a question for that programme’s published terms.
- The instructions for the specific product control
Manufacturer instructions govern installation, and they are product-specific. Exposure, fastener count, fastener placement within the nailing zone, valley method, and how close a joint may fall to a valley centreline are all set by the instructions for the product being installed. A takeoff built on one product’s dimensions and then ordered in another’s is a quantity error and a workmanship exposure at the same time.
Repairability
Order-area discipline has a long tail. A roof ordered with no surplus has no material for the first storm repair, and by then the colour run may have moved. A roof ordered with a deliberate surplus, stored dry and labelled with the lot, has a repair kit. Which of those the customer gets is decided in the takeoff and is worth a line in the proposal — including who owns the leftovers, which is a contract question and not a custom.
What a warranty covers, what voids it, and whether a component substitution matters are set by that manufacturer's warranty document and by the law where the work is. Read the document for the product being installed. This page is not a substitute for one, and it makes no claim about any particular programme.
Six questions to settle before you price itSection link
Not questions for a contractor — questions to settle before you price it. Each one is an item that, left open, becomes a change order or an absorbed cost.
How many layers are on it, and how do you know?
Neither a plan set nor an aerial report can see layer count. A lifted shingle at a rake, a look at the rake edge thickness, or a permit history will. A second layer roughly doubles the weight going into the container and adds materially to tear-off labour, and it is one of the most expensive things a takeoff can be wrong about.
Where does the container go, and how far is it from the ladder?
Carry distance is labour that never appears in a square-foot rate. A container that has to sit on the street rather than the driveway can also mean a permit, a time window, and protection the estimate never priced.
What is the deck, and what is the unit price if it has to come off?
Plank decking, spaced sheathing, or a deck that has already been over-nailed changes the fastener specification and sometimes the whole assembly. Agreeing a unit price and an included quantity before the tear-off starts protects both sides; discovering it on day two protects neither.
Which penetrations are live, which are abandoned, and which are moving?
Abandoned plumbing vents and old satellite mounts are a count nobody makes, and each is a boot, a patch, or a deck repair. Bath fans terminating through the roof rather than the soffit are another. Photograph the roof plane and count from the photograph, not from memory.
Is the ventilation staying as it is?
Ridge vent occupies ridge linear footage and cap goes over it, so they are two quantities and not one. Changing exhaust type, or adding intake, changes both the material list and the labour. Intake and exhaust balance is covered under ventilation.
What is the imagery date on the report, and what has changed since?
A commissioned report measures the roof that existed when the imagery was captured. An addition, a new dormer, a solar array, or a section already re-covered will not be in it. It is a thirty-second question and the cheapest one on this list.
Put these in the proposal, in these words
- Measured surface area and ordered area, stated separately, in squares
- The waste allowance as a number, with the geometry it was derived from
- Linear footage by type: eave, rake, ridge, hip, valley
- Accessory quantities and the specific products they are priced on
- Layers to be removed, and what happens to the price if a third is found
- Deck replacement as a unit price with a stated included quantity, not an open allowance
- Disposal: who provides the container, where it goes, and who pays overweight
- Access and staging: driveway use, protection, and any street or lot permit
Misconceptions and the ways a takeoff actually failsSection link
The ones that survive experience, and the ones that cost the most.
Common misconceptions
Common belief
A steeper roof wastes more material, so it needs a bigger waste factor.
What is actually true
Under the offcut model on this page it does not. Steepness multiplies both the surface area and the length of every cut by the same pitch factor, and the factor cancels out of the percentage. What steepness genuinely adds is breakage, handling, staging and time — real costs that belong in labour and equipment. Putting them in the waste factor hides them and makes the material number wrong in both directions.
Common belief
The aerial report is accurate to within one or two per cent, so the takeoff is settled.
What is actually true
A published accuracy figure is a statement about a benchmark sample, not about your roof and not about your job. The one vendor validation this page could actually read reported 98.77 per cent on roof lines, 98.43 per cent on area and 98.49 per cent on slope, measured by a survey firm against drone and terrestrial LiDAR on single-family houses in the Denver metro in April and May 2025, with the property count not disclosed. That is a real and useful claim about geometry. It says nothing about layer count, deck condition, access, staging, disposal, or what has been built since the imagery was captured — and those are the items that decide whether the job makes money.
Common belief
Starter is an eave product.
What is actually true
The Department of Energy’s asphalt-shingle guide says to install the starter strip at the eave and the rake, and that at rakes it overlaps three inches with the starter at the eave. It says the same about drip edge: eaves and rakes. On Roof A that is 67.1 feet of rake against 80 feet of eave. Ordered off eave length, the starter and drip-edge quantity is 80 feet. Ordered off the takeoff, it is 147.1 — nearly double, on the simplest roof in this article.
Common belief
Cap is cap. Take the ridge and hip footage, divide by the bundle, done.
What is actually true
Divide by which bundle. Within one manufacturer’s published range, a hip-and-ridge product covering 36.5 linear feet per bundle and a high-profile one covering 20 linear feet per bundle both exist. Roof B’s 110 feet of cap is 3.01 bundles of the first, which means buying four, and 5.50 bundles of the second, which means buying six. Same roof, same manufacturer, fifty per cent difference in bundle count, and a takeoff that priced “three bundles of cap” because 110 divided by 36.5 is about three is short before the truck leaves.
Common belief
Waste is padding, so trimming it is free margin.
What is actually true
Offcut waste is a physical consequence of cutting rectangles into diagonals. It is not a contingency and it does not respond to optimism. Trimming the allowance on a cut-up roof does not reduce the material consumed; it moves the cost from the estimate to the second delivery, the crew standing around, and the margin.
How it actually fails
- The linear schedule is produced and never read
- Software reports total squares. The estimator reads total squares. Nobody consumes the linear-footage schedule sitting three lines below it, so hip cap, valley material, rake starter and rake drip edge are estimated from habit rather than from the takeoff.What you can see: Material orders that are consistently right on field shingles and consistently short on accessories. Second trips to the supply house on cut-up roofs and never on gables.
- One waste percentage for every roof
- A percentage calibrated on the company’s ordinary work is applied to everything. On simple roofs it is a silent price increase that loses bids; on cut-up roofs it is a silent discount that wins them. The company then concludes it is good at complex work, because it keeps winning that work.What you can see: A win rate that is markedly higher on complex geometry than on simple roofs, combined with gross margin that goes the other way.
- Measured area and ordered area are treated as one number
- Measured area is priced, ordered area is bought, and the difference is never a line anywhere. It is invisible on a simple roof, where the two numbers are close. On a cut-up roof it is several bundles and the labour to handle them.What you can see: Proposals that state squares without saying whether waste is in the figure. Job costs that miss on material by roughly the waste error, every time, on the same kind of roof.
- Layer count assumed rather than checked
- Tear-off, disposal weight, deck exposure and existing-condition risk scale with layer count, and layer count is invisible to plans, to satellite imagery, and to an aerial report. It is visible from the rake edge in about four seconds.What you can see: Disposal costs that overrun on older housing stock. Change orders raised on day one of a job rather than day three.
- Counts rounded to zero one at a time
- A pipe boot, an abandoned mount, a chimney cricket, a skylight curb: each is cheap, so each is skipped, so the aggregate is never priced. Twelve of them is twelve separate trips across the plane with a different material in hand, and none of that sits in a square-foot rate. This page does not model detail labour; it only insists the count exists before somebody absorbs it.What you can see: Detail work consistently running over its allowed hours while field production runs on schedule.
Sources and further readingSection link
Understanding Roofing / Published
Scope and limitations
- It publishes no dollar figures.
- Takeoff errors are expressed here in bundles, linear feet, fasteners and percentages, because there is no defensible national dataset that would let this site attach a price to a bundle of cap in a particular market on a particular date.
- The money side is at /trade/pricing-and-margin/.
- The offcut waste model is a model.
- It rests on stated geometry, on one explicit behavioural assumption — that the length left in an installer's hand after a cut is effectively uniform across courses and averages half a shingle — and on the assumption that every angled line is cut on both planes, which is true of hips and open valleys but not of closed-cut or woven ones.
- It has not been validated against measured material returns, and no such dataset was found.
- Treat the numbers as a way to compare two roofs, not as a prediction of either.
- It cannot tell you your own waste rate.
- A contractor with job-cost history showing returned and leftover material by roof type has better information than anything here, and should use it.
- It publishes no code determination.
- The 24-inch ice-barrier figure is Department of Energy guidance, not adopted law.
- Requirements vary by adopted edition, local amendment, climate designation and existing assembly, and the authority having jurisdiction governs.
- It publishes no fatality or injury statistic for the trade.
- The Bureau of Labor Statistics Census of Fatal Occupational Injuries would be the right source, and bls.gov returned 403 to every automated request made while writing this page, so no figure from it is asserted here.
- Product figures are product-specific and from two manufacturer sources that could be read.
- Several other manufacturers' data sheets — GAF's in particular — returned 403 and were not used.
- Every coverage figure quoted names its product; substitute the data sheet for whatever is actually being ordered.
29 CFR 1926.501 — Duty to have fall protection
U.S. Occupational Safety and Health Administration
The six-foot fall-protection duty for employees engaged in residential construction activities under 1926.501(b)(13), and that the written-plan alternative under 1926.502(k) is available where another provision of paragraph (b) does not supply one.
A federal regulation. In the twenty-two state plans covering private-sector employment the applicable standard is the state's own, which must be at least as effective but is not always identical. Applicability to any particular company depends on the work, the employment relationship, and the jurisdiction. Nothing here is a compliance determination.
Residential fall protection — compliance guidance
U.S. Occupational Safety and Health Administration / No publication date is printed on the page. Its text refers to compliance directive STD 03-11-002, which OSHA issued on 16 December 2010 with an effective date of 16 June 2011 — the directive is a separate document from this guidance.
That OSHA will accept a properly rigged fall restraint system in place of personal fall arrest where the worker cannot reach the fall hazard, and that a written fall protection plan under 1926.502(k) is the alternative where conventional protection is demonstrably infeasible or creates a greater hazard.
Guidance. The page states that it creates no new legal obligations. It does not determine what any particular employer must do on any particular roof.
OSHA State Plans
U.S. Occupational Safety and Health Administration
That there are 29 OSHA-approved State Plans, 22 of which cover private-sector as well as public-sector workers, and that State Plans must be at least as effective as federal OSHA.
It establishes the structure, not the content. Which standard applies to a given employer, and what it says, has to be read from that state plan.
Asphalt Shingle Roofs — resource guide
U.S. Department of Energy, Building America Solution Center (PNNL)
That starter strip is installed at the eave and the rake and laps three inches where they meet; that drip edge is installed at eaves and rakes; the ice-barrier extent of not less than 24 inches inside the interior plane of the exterior wall in cold climates; that six nails are generally required to fasten one shingle strip and are called for in high-wind areas; and that shingles at open valleys are set in a minimum eight-inch-wide strip of flashing cement.
Best-practice guidance for builders, not adopted law and not a substitute for the installation instructions of a specific product. Its ice-barrier description is guidance for cold climates; the requirement where the work is done is set by the adopted code and the authority having jurisdiction.
IKO shingle dimensions — size and exposure by product
IKO
The shingle length, width and exposure used in the offcut derivation: IKO Cambridge at 40⅞ in × 13¾ in with a 5⅞ in exposure, and the spread of dimensions across products in the same range.
Manufacturer data for one manufacturer's products. The dimensions differ between products in the same table — the Marathon Plus is 39⅜ by 13¼ with a 5⅝ exposure while the Cambridge is 40⅞ by 13¾ with a 5⅞ exposure — which is exactly why a takeoff cannot assume a universal shingle. Substitute the data sheet for the product being ordered.
How many square feet in a bundle of shingles — bundle coverage
IKO
Coverage of 33.3 sq ft per bundle and three bundles per square for IKO Cambridge; and ridge-cap coverage of 36.5 linear feet per bundle for Hip & Ridge 12 against 20 linear feet per bundle for UltraHP — the two figures the cap-bundle comparison on this page rests on.
Manufacturer data, product-specific. The page states plainly that not all shingle brands are packaged three bundles to a square, and the cap figures are for two IKO products only. Do not carry either number to another manufacturer.
EagleView roof measurements confirmed to be 98.77% accurate compared to independent benchmark measurements
EagleView / 4 June 2025
The vendor's own published accuracy figures — 98.77% on roof lines, 98.43% on roof area, 98.49% on roof slope — and the basis for them: testing by CompassData against UAV and terrestrial LiDAR on single-family residences in the Denver metro area in April and May 2025.
A vendor announcement of a vendor-commissioned validation, not an independent study and not a peer-reviewed one. The sample is single-family houses in one metropolitan area over two months and the property count is not disclosed. It supports a claim about measured geometry against a LiDAR benchmark and nothing else: it is not evidence about layer count, deck condition, access, staging, disposal, or currency of the imagery, and it cannot be read as an accuracy guarantee for any individual report.