The site survey checklist, built backwards from rework

Most survey checklists are written by someone who has never had to go back. They list what a survey contains. They do not list what the office will call you about three days later. This one is built backwards from rework: every item is here because leaving it out has sent a technician back to a roof.

Print it, put it on the tablet, or copy it into whatever you already use. The order matters less than the completeness.

1. Before anyone drives

• Confirmed access: gate codes, dog, whether the occupant needs to be present, and a phone number that a human answers.
• Attic access confirmed to exist and be reachable. A survey that cannot get into the attic is a partial survey and the office should know before the truck leaves.
• Panel location known, or at least whether it is interior or exterior. This changes what gets loaded on the truck.
• Roof material and storey count. A tile roof and a two-storey walk are different equipment and different time.

2. The roof

• Every plane, photographed square-on and from a raking angle. The raking shot is what shows deflection and surface condition.
• Ridge, eave, and rake measurements per plane. Measured, not paced.
• Azimuth and pitch per plane.
• Rafter or truss size, spacing, and direction, from inside the attic. This is the structural input and it cannot be guessed from the exterior.
• Sheathing type and thickness, plus any delamination or moisture staining.
• Layer count at a cut edge or vent penetration. Two layers of comp changes the attachment and sometimes the answer.
• Every obstruction with a measurement to the nearest plane edge: vents, stacks, skylights, chimneys, satellite mounts, swamp cooler, existing conduit.
• Condition note with a year estimate. A roof with three years left is a conversation, not a mount.

3. Shading

• A full horizon capture from each array area, not one from the middle of the roof.
• Trees photographed with something in frame for scale, and noted as on-property or neighbour-owned. Whose tree it is decides whether trimming is an option.
• Near obstructions that will shade in winter and not in summer: the chimney, the second storey, the neighbour's gable.

4. Electrical

• Main breaker rating and bus bar rating, photographed as two separate labels. They are often different and the bus rating is the constraint.
• Deadfront off, full interior photo, with every breaker legible.
• Slot count: used, open, tandem, and double-tapped.
• Meter number and meter-main versus separate configuration.
• Service entrance: overhead or underground, conductor size, and route.
• Grounding electrode and bonding path.
• Utility and rate schedule if visible on the meter or a bill.
• Proposed inverter and conduit route, with a measured run length along the actual path.

5. Site and staging

• Where the crew parks and where material stages, photographed.
• Ladder setpoint options, and whether any of them are on a slope.
• Overhead service drop clearance if the ladder or array is near it.
• Anything that will be in the way on install day: RV, boat, low branches, fence with no gate wide enough.

6. Before leaving the site

This is the part most checklists skip, and it is the part that decides whether the survey was any good.

• Every photo reviewed on the device, not on the drive home. Blurry, dark, and cropped-too-tight are all recoverable while you are standing there and expensive afterward.
• Every measurement written down, not held in memory for later entry.
• Upload attempted on site. If there is no signal, that is a note on the job, not a surprise for the office.
• One sentence in plain language about anything that felt wrong. The technician's instinct is real data and there is usually nowhere on the form for it.

Why a checklist beats a good technician

We looked at this directly in two technicians, two roofs: two people on the same pay, doing the same number of jobs, with ten points of difference in rework rate. The gap was not skill or care. It was that one of them was working from a spec and the other was working from experience, and experience varies by how tired you are at four in the afternoon.

A checklist converts a judgement call into a yes or no. That is the only reason a flat price works. We run this spec across more than 50,000 surveys over seven years, nationwide, at a flat $275 per site, with the deliverable back inside 48 hours and a 3D roof model accurate to less than an inch. None of that is possible if the capture is a matter of opinion.

If you are weighing whether to run this in house or send it out, the arithmetic is in the algorithm no one is running, and the reason the number moves is in utilization.