Ask a contractor how they are doing on bidding and you will usually get a hit rate. One in four, one in six, one in ten depending on the market and the work. It is a comfortable number to talk about because it is a number about competition, and competition is somebody else's fault.
The number nobody quotes is how many invitations came in that never got priced at all.
That one is uncomfortable because it is not about competition. It is about capacity. And in most contracting businesses, the thing that runs out first is not sales, not bonding, not crews. It is the hours available to measure a drawing set.
What takeoff actually costs, hour by hour
Takeoff is the step where somebody measures quantities off a set of drawings so the work can be priced. Linear feet of pipe. Square footage of slab. Counts of fixtures. Cubic yards of cut and fill. It happens before pricing, and it governs everything after it — because if the quantity is wrong, the bid is wrong no matter how good your unit costs are.
The cost of doing it by hand breaks into four parts, and only one of them is the measuring.
Finding what matters. A commercial set can run to hundreds of sheets. Before a single quantity is measured, an estimator has to work out which sheets carry the scope, which details govern, and which of the three apparently contradictory notes is the current one. On a large set this is hours of work that produces no output at all.
Measuring. This is the part everyone pictures — clicking along a wall on screen, or scaling off a printed sheet, and typing numbers into a spreadsheet. It is slow, it is repetitive, and it is exactly the kind of task where attention degrades over a long afternoon.
Re-keying. The quantities then get typed into the estimating package where the assemblies and labour rates live. Every re-key is a second opportunity for a digit to move, and it happens under deadline.
Re-doing it when the addendum lands. Which it does, usually two days out. Someone works out what changed, then re-measures the affected scope. In practice this often means measuring more than was strictly affected, because there is no reliable record of what the original numbers covered.
Add those up and the pattern is clear enough: the majority of takeoff time is not spent measuring. It is spent finding, transcribing, and redoing.
The error nobody catches
There is a second cost that is harder to see, and it is not arithmetic.
When a quantity is produced by hand, there is usually no record of where it came from. The spreadsheet says 1,840 linear feet. It does not say which sheets, which walls, or which assumptions about what counts as in scope. So when someone wants to check it, checking means re-measuring — which takes as long as measuring did.
The practical result is that most numbers are never checked. Not because estimators are careless, but because checking costs the same as producing, and the deadline only funds one of those.
This is why takeoff errors tend to surface at the worst possible moment: not during review, but during construction, when the installed quantity does not match the bid quantity and nobody can reconstruct which one was wrong.
The bids you never submitted
Here is the part that actually costs money.
If takeoff capacity is fixed — and it is, because it is a function of how many experienced estimators you have and how many hours they have — then bid volume is capped. When more invitations arrive than there is capacity to price, somebody triages. Usually informally, usually under pressure, usually by picking the ones that look most winnable.
Every invitation that gets declined for capacity reasons is a bid with a zero percent hit rate. It does not show up in the hit rate calculation, because it was never a bid. It shows up as revenue that never existed, attributed to nothing.
This is the number worth putting a figure on. Not "we lost that job", but "we were invited to price forty jobs last quarter and priced twenty-six." The fourteen are invisible in every report you run, and they are the largest single lever in the business.
Why hiring is not the obvious answer
The reflexive fix is to hire another estimator. Sometimes that is right. But it is worth being clear about what you are buying, because takeoff time is not homogeneous.
Roughly speaking, takeoff work splits into judgement and measurement. Judgement is knowing which detail governs, what the drawings do not say, where the scope gaps between two trades sit, and what this particular owner tends to argue about later. That is expensive to hire, slow to develop, and genuinely scarce.
Measurement is counting things accurately and consistently.
When you hire an estimator to increase bid capacity, you are paying for judgement and using a large fraction of that person's week on measurement. That is not a criticism of how anyone runs their business — it is just how the work has always been packaged. But it does mean that increasing capacity by hiring is expensive per unit of capacity gained.
The alternative is to reduce the measurement load on the people you already have, so more of their week goes to the part only they can do.
What changes when the measurement is proposed for you
This is what AI takeoff is actually for, and it is worth being precise about it, because the category has attracted more enthusiasm than accuracy.
Software reads the drawing set and proposes quantities for the repetitive, measurable scope. Each proposed quantity is linked back to the sheet region it came from. The estimator reviews and corrects rather than producing every number from scratch.
Three things follow from that, and they matter in a specific order.
The first is throughput. Review is faster than production, so more sets get priced in the same week.
The second is checkability. Because every number traces back to a drawing region, checking a quantity takes seconds instead of being a re-measurement. So checking actually happens, which changes the error profile more than the speed does.
The third is addendum handling. A revised set can be compared against the measured one, so only affected scope gets re-reviewed. The two-days-out addendum stops being an all-nighter.
What does not change: judgement. Ambiguous scope, contradictory details, and the gaps between trades still need an estimator, and any system that claims otherwise is one you should not trust with a bid. Good tooling flags those rather than guessing at them.
Where this shows up differently by trade
The measurement problem is universal in construction, but it is not the same problem everywhere, and the tooling only helps if it is built for what you actually quantify.
A commercial general contractor is measuring assemblies across every trade in the set and worrying about the scope splits that decide who bids what. A site work contractor is computing cut and fill between surfaces, where a swell and shrink assumption buried in a spreadsheet is the difference between a good job and a bad one. A production builder is not measuring jobs at all — they are maintaining plan-level bills of material with option variants, built forty times over. A remodeler is measuring existing field conditions rather than drawings.
Same category, four different jobs. Which is why generic takeoff tooling tends to help one of them and irritate the other three.
The number to actually go and find
If you take one thing from this: go and count the invitations you did not price last quarter.
Not the bids you lost. The ones that arrived and were declined, deferred, or quietly ignored because there was no capacity. Then multiply by your average contract value and your hit rate.
That number is the cost of manual takeoff. Everything else in this article is detail.