BESS installation labour estimator

Why it ships without labour units

The obvious way to build this is to load a published labour manual into the page. That is not available to us, and the reason is worth stating plainly rather than hiding behind a disclaimer.

Manuals like NECA’s are commercial publications, and their labour-unit tables are the product being sold. Owning a copy licenses you to use it — including in your own estimating, on your own machine. It does not license anyone to republish the tables on a free web page. And the numbers are not raw facts that copyright leaves alone: they are predicted values reflecting editorial judgement, which courts have treated as protectable in exactly this shape.

The alternative — inventing plausible-looking units so the tool shows a total on landing — is worse. The stated use of this page is checking a contractor’s number. A confident total built on made-up productivity would be the most damaging thing it could do.

So the page ships 316 unit records with their structure fully defined — category, size, install method, unit basis, assumed crew, whether condition changes them, whether the termination already includes torque — and no values. An unpriced row says not priced. It never contributes a zero.

Getting to a number

1. Pick a block and write the cable schedule. Enclosure count and power, then a line per scope: how many runs, how far, how many conductors, what size. The takeoff follows the schedule, so every cable row on the sheet is one you wrote and can correct — and the per-enclosure scopes (equipment setting, comms, conduit, tray) re-derive from the enclosure count.
2. The schedule does not multiply itself. Runs and lengths are what you measured, so changing the enclosure count leaves them alone. If a bigger block means more cable, that is a number only you can state.
3. Type the unit values, in the panel under “Labour units”. This takeoff needs 28 of the 316 records — not all 316 — and they are listed in the takeoff’s own order, each showing the basis, the assumed crew and whether the value is man-hours or crew-elapsed hours. Values come from a source you are entitled to use: your licensed manual, your estimating database, or the hours your own crews book against these tasks. The last of those is the best input anyone has.
4. Or bring them as a file. Download the 28 units this takeoff needs (or all 316), fill the sheet, and paste or upload it back. Values stay in your browser and are remembered between visits. They are not transmitted to us, and clearing them is one button.
5. Enter one wage rate. Cost needs a base and a fringe for each craft the takeoff uses; one composite rate can fill them all, and each field it fills says so. Until a rate exists the ledger returns nothing rather than guessing.

The sheet records each unit’s basis and crew. If a unit’s definition changes after you exported the sheet, the import refuses that value rather than applying it — the same number under a different basis is a different number.

The four things this gets right that spreadsheets usually do not

• Polarity is a conductor, not a runA DC pull carrying + and − is one pull with two conductors. Cable units are per conductor-foot, so treating the pair as two runs doubles the largest line on the sheet. The generator decides this once and the row echoes the basis it resolved.
• Condition selects a column, it does not multiplyNormal, Difficult and Very difficult are three independent values. Nothing here derives the harder tiers from the easy one by a fixed ladder. If you only have a Normal value, you can apply your own step — and it is labelled as yours.
• Torque is a property of the termination, not a separate multiplierIf a termination unit already contains torque-to-spec and witness marking, no separate torque row is generated for those joints. Put both in and you have counted the joint twice; the engine refuses that outright and names both rows.
• One clockAttended hours per shift is the only divisor for duration and man-days. Site efficiency multiplies hours, once. Dividing paid hours by a net-of-breaks figure and then also applying an efficiency factor takes the same haircut twice, and it is a common way per-diem estimates come out 50% high.

Limits worth knowing before you rely on it

It prices direct installation man-hours first, then adds indirect supervision — foreman by ratio with a minimum-foreman floor, and general foreman by percentage. Those ride in cost and in the total and never in crew-days, because supervision does not shorten a job. The ledger also carries a safety attendant on any row marked as an energised work front, but nothing on this page can mark one yet — no shipped record is flagged and the unit sheet has no column for it — so that term is always zero today.

Cost is optional, and it is withheld rather than guessed. No wage rates ship, for the same reason no labour units ship: you enter your own, and until every craft carrying hours has one the ledger returns nothing. It is withheld while any row is unpriced, and while prevailing wage is switched on without a multiplier — there is no safe default for that number, so the ledger refuses rather than assuming one. The switch lives under “Cost details” in the rates section.

What it still does not model: escalation to a construction midpoint, shift differential, a detailed equipment schedule, per-craft workers’-comp rates, cash-in-lieu fringe, or anything below the line — bonding, insurance, permits, sales tax on materials. Per diem, mobilisation and demobilisation, travel, temporary facilities, the overtime and double-time split and the OEM field-service day rate have inputs in the rates section, most under “Cost details”; anything left at zero is listed as zero rather than hidden.

Per-MWh and per-enclosure figures are withheld whenever any row is unpriced. A normalised number computed over an incomplete scope is precise, quotable and wrong, and it is the failure mode most likely to end up in somebody’s board pack.

The learning curve is implemented and structurally gated: it applies only to units that are first-unit values, and no record in this dataset is one — so it could take no repetition saving even if it were switched on, and no control offers it. That is the safe direction to be wrong in: applying a cumulative-average curve to a unit that is already an average takes roughly 28% off at a hundred repetitions, which is a large and entirely fictional saving.