The bid is due Thursday at 2pm. Your estimator has been at the desk since six, and by the end of the day the company will have committed to a price on a job worth more than most houses. Watch how those eight hours actually get spent, and the problem becomes hard to unsee.
Thursday, hour by hour
6:10am — coffee, plan sets open. Addendum 3 dropped at 4:50 yesterday afternoon and moved a wall, which means part of Tuesday's take-off is now wrong. The first hour goes to re-measuring quantities that were already measured once.
8:30 — the price book. Every line of take-off — linear feet of this, square feet of that, counts of the other — needs a unit cost set against it. Some costs come straight from the book. Some come from a supplier quote that expired last week and needs a phone call to refresh.
10:00 — chasing subs. The mechanical number was promised for Wednesday. The electrical sub picks up on the third try and says his number is coming.
11:20 — it arrives: a photo, taken in a truck cab, of the spec sheet with "$48,200 all in" written across the corner in pencil. No scope letter. Does "all in" include the trenching for the service run, or is trenching excluded the way it turned out to be excluded on the last job — discovered after award? That question is another phone call.
12:40 — retyping. Numbers from photos, PDFs, emails, and one voicemail, keyed by hand into the bid sheet. This is the hour in which a transposed digit becomes a five-figure mistake, and everyone in the office knows it, which is why nobody talks to the estimator at 12:40.
1:15 — the proposal. Scope, exclusions, alternates, unit prices where the bid form demands them, the form itself filled out exactly the way this owner requires, the bid bond attached.
1:56 — submitted. Four minutes to spare, which around here counts as a calm week.
Two different jobs wearing one title
Run that day back and it splits cleanly in two. One job is judgment: deciding what this work will actually cost your company, with your crews, on that site, for that customer. The other job is collation: gathering numbers from a dozen sources in a dozen formats and moving them into documents without breaking anything. Your estimator was hired for the first job. Thursday went almost entirely to the second.
What the judgment actually is
Nobody automates the part that wins work, and nobody should claim to. Judgment is knowing that trenching west of town runs slow because you hit caliche two feet down. It is knowing this general contractor pays in ninety days, and the price has to carry that float. It is reading the construction schedule in the specs, recognizing it as fantasy, and planning for a crew that mobilizes twice. It is finding the spot where the drawings and the spec book disagree and deciding whether to price the conflict, exclude it in writing, or send the RFI before bid day. That knowledge took fifteen years of built work to accumulate, and it is the actual reason your hit rate is what it is.
What the collation actually is
The rest of Thursday is transport. Quantities travel from the take-off into the pricing sheet. Unit costs travel from the price book and supplier quotes onto the quantities. Sub numbers travel from a pencil-marked photo and three email threads into line items, and the totals travel into the proposal template. Every hop is manual, every hop can drop or flip a digit, and not one hop uses the scarce skill you pay for. It is clerical work performed at expert wages, under a deadline, with the company's margin riding on the typing.
The half a machine can carry
Structured intake changes the shape of the day. A sub's quote — the photo, the PDF, the email — lands in one place, gets read by software, and comes out as data: bidder, trade, amount, what is included, and what is silent. The silence is the important part. If the electrical quote never mentions trenching, the system flags the hole instead of letting "all in" hide it until after award. Price-book lookups become retrieval instead of memory. The proposal assembles itself from the templates the estimator already trusts, with numbers flowing from a single source rather than being retyped at every stop. The estimator's role at each stage shifts from transcribing to reviewing — which is faster, and which is also the posture in which humans actually catch errors.
What this does not do
It cannot price risk. No system knows that the GC pays slow or that the soil is wrong; it prices the job in front of it, and the job in front of it is never quite the job that gets built. It cannot capture what happened on a phone call — if the sub verbally excluded the trenching and wrote nothing down, there is nothing to extract. A spec sheet photographed at a bad angle in bad light can read wrong, which is why extracted numbers surface for a human's confirmation instead of flowing silently into totals. And a price book that has not been touched since materials moved is automated garbage: the system faithfully multiplies whatever it is given. Cleaning the book is part of this work, not a footnote to it.
What changes on bid day
The deadline does not move and the judgment does not shrink. What changes is the proportion. The re-measure after an addendum still takes the hour it takes. But the sub chase becomes a status board instead of a call sheet, the 12:40 retyping hour mostly disappears, and the afternoon belongs to the questions that decide whether you win at a price you can live with. It also leaves room for the thing most shops never do: the follow-up call on last week's outstanding bids, which dies every Friday because everyone is already buried in the next one.
Where to start
Ask your estimator to keep a plain time log through one full bid cycle — two columns, judgment or collation, in half-hour blocks. No software, just a legal pad on the desk. Most owners who see the result find the split is worse than they guessed. That log, not anyone's pitch, is the case for changing how Thursday works, and it is the first thing we would ask to see anyway.