A construction takeoff is the process of measuring and counting everything a job will require — the material quantities, linear footage, surface areas, and unit counts — and pulling those numbers off a set of drawings or out of a physical building so they can be priced. The name comes from the old practice of "taking off" quantities from the plans. Every bid a contractor submits sits on top of a takeoff, whether the contractor did it deliberately or guessed.
The takeoff answers one question: how much of each thing does this job need? How many square feet of TPO membrane, how many linear feet of EMT conduit, how many devices on the fire-alarm riser, how many modules on the array, how many tons of RTU. It is the quantity layer. It is not yet the money layer. Confusing the two is the most common reason a contractor's numbers drift, so it is worth being precise about where the takeoff ends and the estimate begins.
Takeoff vs. estimate — the distinction that matters
A takeoff produces quantities. An estimate applies costs to those quantities — labor rates, material pricing, waste factors, equipment, overhead, and margin — to produce a number you can put on a proposal. The takeoff is the input; the estimate is the output. You cannot have a defensible estimate without a takeoff under it, and a perfect estimating spreadsheet built on a sloppy takeoff is just a precise wrong answer.
The reason the distinction matters operationally is that the two failures look different and cost different. A takeoff error — counting 40 rooftop penetrations when there are 60, or measuring a roof at the parapet line instead of the membrane line — propagates through every downstream calculation and is usually invisible until the crew runs short on material in week three. An estimating error — a stale unit price, a labor rate that didn't account for prevailing wage — is bounded and easier to catch. Most contractors who lose money on a job lose it at the takeoff, not the estimate, because the takeoff is the step that gets rushed.
Commercial installation work makes this worse than residential, because the assemblies are denser and the quantities interact. A single rooftop unit isn't one line — it's the curb, the flashing, the membrane termination around it, the conduit feeding it, and the labor to integrate all of it. A takeoff that counts the unit but misses the assembly around it understates the job by the part that actually takes the time.
How takeoffs get done today
There are three broad methods in practice, and most contractors use more than one depending on the job.
Manual takeoff from paper or PDF plans. The estimator prints the drawings or opens them in a PDF markup tool — Bluebeam Revu is the category standard — and walks the sheets with a digitizer, a scale, and a highlighter, tallying counts and measuring lengths and areas by hand. It is slow, it is accurate in skilled hands, and it lives or dies on the quality and currency of the drawings. On a tenant-improvement or retrofit job where the as-builts don't match the building, the plan-based takeoff measures a building that no longer exists.
Digital plan takeoff. Software like PlanSwift, STACK, or a trade-specific package overlays measurement tools on the drawing set and auto-totals quantities as the estimator marks them up. Faster than paper, with running tallies and exportable quantity lists. Still fundamentally a drawings-based method — it inherits whatever the plans got right and whatever they got wrong.
Field measurement. The estimator measures the actual building — historically with a tape and a wheel and a laser distance meter, increasingly with 3D capture. This is the only method that measures ground-truth: the building as it stands today, not the building the architect drew or the surveyor flew. For re-roof, retrofit, and any work where the existing conditions drive the scope, field measurement is the method that doesn't lie. The cost has always been time — a careful manual field takeoff of a commercial roof is a half-day with a clipboard.
Aerial imagery — a fourth method, and its limits
A fourth category sits between plans and field: aerial-imagery measurement. Providers like Nearmap, EagleView, and GAF QuickMeasure reconstruct a roof or site from satellite or drone photography and deliver a measurement report, often within about a business day. For a contractor who can't safely or quickly get on a roof, an aerial report is genuinely useful, and for some residential and storm-response workflows it is the right tool. It deserves credit where it earns it.
It carries two structural limits worth naming honestly. First, the imagery is only as current as the last flyover, so a building that was reroofed or retrofitted since the last pass measures the old conditions. Second, photogrammetry from above can't see what a parapet hides, what an overhang covers, or what sits inside a courtyard; the curbed equipment, the ballast depth, the penetration field — exactly the geometry commercial estimators care about — is the geometry aerial struggles to capture cleanly. On flat, unobstructed, recently-flown roofs it does well. On the messy commercial geometry that drives real scope, it's an approximation delivered tomorrow rather than a measurement taken today.
Why the takeoff is the leverage point
The takeoff is the single highest-leverage step in the whole bidding chain because every other number depends on it and because it's the step under the most schedule pressure. Bids have deadlines; the takeoff is the part that gets compressed when the deadline is close. An estimator with three bids due Friday does fast, conservative takeoffs and pads the number to cover the uncertainty — which means the contractor is either leaving margin on the table by over-padding or buying the job by under-counting and eating the gap in the field.
There's a second cost that doesn't show up on any one job: the takeoff is where institutional knowledge leaks. The senior estimator knows that this membrane needs a 12% waste factor and that this AHJ requires a particular smoke-detector spacing, and that knowledge lives in their head and their personal spreadsheet. When they take a day off, or leave, the takeoffs get worse, and nobody can see why until the margins slip. A takeoff method that captures that knowledge in the system instead of the estimator's head is worth more than the time it saves.
This is the reason takeoff and estimating are converging into one motion. The traditional split — measure first, price later, in two different tools — exists because the tools were built separately, not because the work is naturally two steps. When the quantity you mark on the plan immediately carries its assembly, its waste factor, and its current cost, the takeoff and the estimate stop being two jobs.
How Forge handles takeoff
Forge collapses takeoff and estimate into one motion, and it does it from the strongest input — the building itself. Forge's estimating runs on trade assemblies: when a quantity is captured, it carries the full assembly behind it, not just the bare count. Capture the rooftop unit and you capture the curb, the flashing, the termination, and the labor — because that's how the assembly is defined, the way a commercial estimator actually thinks about scope.
The ground-truth input is Hyperion — Forge's iPad and iPhone LiDAR scanning system. Instead of measuring stale plans or a months-old flyover, the operator walks the structure and the sensor captures the spatial geometry of the building as it stands. From that capture, the federation kernel pulls the assemblies, waste factors, and material costs, and a line-item estimate is generated. The takeoff and the priced estimate come from the same pass — off the real building, on site, before the operator leaves the roof.
On accuracy, Forge stays honest about where the number comes from. Forge publishes ±0.8% on commercial flat and low-slope geometry, widening to ±2% on tighter, more complex geometry. That figure is currently published as a documented methodology and sample plan at /proof/hyperion-accuracy — it graduates to a measured benchmark after independent third-party verification, which is not yet complete. The proof page tracks updates with dates so a reader can see exactly what's been established and what hasn't. We don't claim it as already-independently-verified; we cite it the way the proof page does.
Hyperion is built first for commercial roofing, and it ships today across the verticals Forge serves — roofing, security and fire, AV and low-voltage, solar, HVAC, and electrical. It's a $399/mo module on Forge Core, included in the Working Stack and Full Platform packages and in the Charter founder program — not sold as a separate line item. The point isn't a faster digitizer. The point is that the takeoff stops being a separate, error-prone, knowledge-leaking step performed against the wrong copy of the building, and becomes the moment you walk the job — measured against ground-truth, priced by the assembly catalog, captured in the chest instead of someone's head.