An estimate is a quantity multiplied by a rate. The quantity take-off gets all the attention. The rate does all the damage.
Quantities are auditable. You can measure them off drawings, reconcile them against the model, and defend them line by line. Production rates are where judgement enters the estimate, and judgement is where estimates fail. On capital projects, a 15% error in a handful of governing rates will do more harm than a 5% error across the entire take-off, because the rate error compounds through manhours, crew sizing, plant schedules, indirects and time-related costs.
First principles or nothing
There are two ways to price work. Estimating by analogy takes what the last project cost and adjusts it. First-principles estimating builds each line up from crew composition, plant, materials and a production rate: the output a defined crew achieves per shift under stated conditions.
Analogy is fast, and it is defensible right up until someone asks a hard question. Which project? Which conditions? What changed? First-principles estimating survives those questions because every number traces back to a rate, a crew and an assumption that can be examined and challenged.
That places the entire weight of the estimate on one input. Get the rate right and the manhours, the histogram, the plant schedule and the cost all follow. Get it wrong and no amount of contingency fixes it.
Where generic rate books fail
The rate books most estimators reach for were calibrated in North American and northern hemisphere industrial corridors: dense contractor markets, established labour pools, short supply lines, temperate conditions. Those rates are honest for the conditions they describe. The problem is applying them elsewhere.
A Gulf Coast piping rate does not apply to a processing plant in Limpopo. The differences are not marginal:
Altitude. Much of Southern Africa's mining work sits above 1,200 metres. Diesel plant derates. People derate. Published rates rarely say at what elevation they were measured.
Climate. A Highveld wet season, or 40-degree heat in the Kalahari, changes achievable outputs and working windows in ways a book rate never captures.
Labour market and crew composition. Crew mixes, supervision ratios, skills availability and local content requirements differ fundamentally from the markets where book rates were measured. The rate is a property of the crew, not just the task.
Logistics. Remote, camp-based sites with long supply lines lose productive hours to travel, materials availability and equipment support. On a remote site, the constraint is often not how fast a crew can work but how consistently it can be fed with work.
Apply an unadjusted book rate under these conditions and the estimate is optimistic before the first excavator arrives. The schedule inherits the optimism. Then the claims begin.
The inherited spreadsheet is not a rate library
Most estimating teams know this, which is why they do not rely on the book. They rely on the spreadsheet. Every estimator we have worked with has inherited one. Nobody remembers who built it. Nobody knows where the rates came from. But it works, mostly, so everyone keeps using it.
The problem is not the spreadsheet. The problem is what happens when you need to price a 500-line bill of quantities across a dozen trades, justify a manhour estimate to a client who wants an audit trail, apply location factors for a site you have never worked, or hand the estimate to a planner who needs trade-allocated resource loading rather than a headline number. That is where it breaks down. Not because the rates are wrong, but because there is no structure, no version control, and no way to trace a number back to its source.
What a defensible rate looks like
A production rate you can stand behind in a review, or in a dispute, has four properties.
It has provenance: you know which projects and conditions it came from. It has structure: discipline, unit of measure, crew composition and stated base conditions, so it can be compared like for like. Its adjustments are explicit: altitude, climate, access and market factors are applied visibly, not baked invisibly into the number. And it is versioned: when a rate is recalibrated, the history survives.
This is why we built our rate library the slow way, calibrated from delivered projects rather than copied from publications. It now covers more than 15,500 validated production rates across 35 disciplines, with the regional and site-condition context attached to each. When we tell a client a duration or a manhour budget is realistic, that is the evidence behind the statement.
The owner's test
If you are reviewing an estimate rather than building one, three questions expose the rate risk quickly. Where did the governing rates come from? What conditions do they assume, and how do those compare to this site? And what adjustment factors were applied, by whom, on what basis? An estimator with a real rate library answers in minutes. An estimator with an inherited spreadsheet cannot answer at all.
The rate is the estimate. Everything else is arithmetic.
ProdRates puts the full 15,500-plus-rate library behind your estimating team, with every calculation auditable from BOQ line to manhour. Visit the ProdRates product page to request a demo, or contact us.