Landscape estimating and construction software

A takeoff is the step where a design or a site visit becomes a list of quantities. Square feet of paving. Cubic yards of base, topsoil and mulch. Linear feet of edging, wall block and irrigation pipe. Counts of plants by size. Everything that will be bought, in the units the supplier sells it in.

The estimate is the step after. It attaches a price to each quantity, adds labour hours, adds equipment and haulage, then adds overhead and margin to arrive at a number the customer sees.

Landscape takeoff software exists to do the first step from a drawing or an aerial image instead of a tape measure and graph paper, and to hand the result straight to the second step without anyone retyping it. That handoff is the whole value. Retyping is where a 9.26 becomes a 9.2, and a job loses a yard of stone nobody ever accounts for.

AI Landscape Design does neither step. It has no measurements in it at all.

A building takeoff mostly deals in rectangles. A landscape takeoff deals in curves, and the difference is expensive.

Take a kidney-shaped bed 24 feet long. Measure it as the rectangle it fits inside — 24 by 11 feet — and you get 264 square feet. Measure it properly, by taking the width every 4 feet along its length and averaging, and six readings of 6, 9, 11, 10, 7 and 4 feet give an average width of 7.83 feet. That is 188 square feet. The bounding rectangle is 40% over.

Forty per cent over on one bed is a rounding error. Forty per cent over on every curved bed, edge and path in a whole-yard job, priced against a competitor who measured properly, is how you lose the bid. Forty per cent under, from a contractor guessing low to win it, is how you lose the job's margin instead. The strip-and-average method takes four minutes per bed and removes both.

Every bulk material on a landscape job is priced by volume and specified by depth, and the conversion trips people constantly because it crosses units.

A 1,000 square foot bed mulched 3 inches deep: 3 inches is 0.25 feet, so 1,000 × 0.25 = 250 cubic feet. A cubic yard is 27 cubic feet, so that is 9.26 cubic yards.

Now change the spec to 4 inches. The same bed becomes 333 cubic feet, or 12.35 cubic yards. One inch of extra depth is 33% more material on the same area, and one inch is exactly the kind of detail that gets agreed verbally on site and never makes it back to whoever priced the job.

This is the single most useful thing estimating software does. Enter area and depth, get a delivered volume, and have the number recalculate when the spec changes rather than being rebuilt by hand. Landscape construction management software that carries an assembly library does the same job for a full patio build-up in one line.

Excavated soil takes up more room out of the ground than it did in it. Dig a 400 square foot patio 10 inches deep and you have removed 333 cubic feet, or 12.35 cubic yards, of ground. Loose in a pile or a truck bed, at a swell factor of around 25% for typical topsoil and clay, that same material occupies roughly 15.43 cubic yards.

You pay to haul and tip the 15.43, not the 12.35. Estimating off the in-place figure understates disposal by about a quarter on every dig, and disposal is a cost that does not shrink when a job goes badly — it is one of the few line items on a landscape job that is genuinely fixed by physics.

Swell is also why spoil never fits back where it came from. A contractor who plans to regrade the excavated material into a bank somewhere on site has to plan for a bank a quarter larger than the hole, or a truck.

Aggregate goes the other way from spoil. It arrives loose and ends up compacted, and the spec is always written for the compacted state.

A 400 square foot patio on a 6 inch compacted base needs 200 cubic feet, or 7.41 cubic yards, of finished base. Crushed aggregate commonly loses about 20% of its loose volume under a plate compactor, so to finish 7.41 compacted yards you order 9.26 loose yards. Order 7.41 and the crew runs out at roughly four fifths of the way across, on a day when the paving is already on a pallet in the driveway.

The trap is that spoil and base pull in opposite directions and both get called "the volume". One is a quarter bigger than the hole, the other is a fifth smaller than the delivery. Guessing either from the raw geometry is wrong twice. Software that carries a material library with swell and compaction factors already attached is worth its subscription for this alone.

Landscaping purchasing software is the layer between the takeoff and the supplier: turning quantities into orders, holding supplier price lists, and tracking what was actually delivered against what was priced.

The first thing it stops is buying in the wrong format. A cubic yard is 27 cubic feet, which is exactly 13.5 of the two-cubic-foot bags sold at retail. That 1,000 square foot bed at 3 inches is either 9.26 yards delivered loose or 125 bags loaded by hand — same material, wildly different money and an afternoon of labour either way. Bagged is right for a courtyard with no truck access and wrong for almost everything else.

The second is stale pricing. Aggregate, mulch and plant costs move between seasons, and an estimate built on a price list from last spring is quietly wrong on every line. The third is delivery and minimum-load charges, which are real money on a small job and are the line most often left out entirely.

Every material on a landscape job gets ordered above the measured quantity, because cuts, breakage and uneven ground all consume material the drawing does not show.

The common allowances: 5% on paving laid in straight runs, 10% where the layout is diagonal or the edges curve, 10% on bulk aggregate and topsoil, and a plant count taken as measured because nurseries take returns and stone yards mostly do not. On a 400 square foot terrace with a curved edge, 10% means ordering 440 square feet.

The rule is right. What goes wrong is that nobody tracks the other side of it. Ten per cent over on every job, never returned and never counted back, becomes a corner of the yard filling with part pallets and half bags — material the business has already paid for and will never invoice. Job costing catches it if the surplus is booked back against the job. A spreadsheet estimate almost never does.

Published pricing in this category starts at $297 a month. LMN's Starter plan is $297 billed monthly for one office and five crew licences and sells budget-led estimating as its core idea; Professional is $648 a month and adds real-time job costing against the estimate. Jobber is the other name every contractor will mention, and it carries quoting rather than true takeoff — no price for it here, because its pricing page sat behind a bot challenge and served us nothing on August 3, 2026.

Above that sit the design packages that produce takeoffs as a by-product of drawing: DynaScape Design at $1,599 a year, Structure Studios Vip3D at $2,004 a year for one seat.

All of it is built for a company with crews. Roughly 78% of US landscaping businesses are sole proprietors with no employees, and a well-built spreadsheet with correct depth and swell factors prices their work perfectly well. Our contractor software guide covers that end.

A perfect estimate loses to a worse one all the time, because the customer is not comparing arithmetic. They are comparing two documents, and a homeowner cannot audit either one. What they can do is decide which contractor seems to understand what they want.

That is a presentation problem, not an estimating problem. The proposal that wins is usually the one that puts a picture of the customer's own yard next to the number, so the price is attached to something they can see rather than to a paragraph of specification they have to imagine.

AI Landscape Design, published by Skyble Lab, does exactly that one thing. Photograph the yard, pick from 18 styles, and get the same yard redesigned in about 10 seconds, then attach it to the proposal. It is $59.99 a year or $7.99 a week, and the first 3 designs are free at full resolution. It produces no quantities, no depths, no costs and no drawing — it will not tell you a single number that belongs in the estimate.