Six numbers, all pulled from jobs you have already built, tell you whether the next ninety days will hurt: minutes to first contact, days to a sendable quote, whether the estimate was opened, win rate split by job shape, days from a finished milestone to money in the bank, and estimate against actual on three line items. Friday's bank balance tells you where you have been, eleven weeks late. There is no industry-average table here, because a benchmark is worth exactly what its definitions are worth — and job tracking software for construction only earns its keep when it timestamps the events those definitions depend on, so the numbers fall out of running the job instead of a spreadsheet somebody has to maintain.
Why is there no benchmark table?
Three things break one. Sample composition: twelve residential garages a year and forty ag buildings are not the same business, and a blended median describes nobody. Definition drift: two shops both report win rate, one counting every phone call and the other only priced, delivered estimates — 21% and 68% off the same year of work. Seasonality and survivorship: a February median and an August median are different numbers from the same shop, and everyone inside an aggregate is still in business. For industry context the National Frame Building Association is the source; the U.S. Small Business Administration covers the financial side plainly.
Number one: how many minutes to first contact?
The minutes between a lead landing and a human saying something back. A homeowner pricing a 40x60 shop calls three builders the same evening, and the first to talk to them sets the frame — eave height, truss span, concrete in or out — that every later quote is read against.
Measure it from your last thirty leads: time arrived, time of first outbound, take the median. What moves it is routing, not effort — who owns the phone between four and eight in the evening, and what happens to a web form at 9pm Saturday. Read the distribution: a twenty-minute median with 30% of leads past twenty-four hours means you are fast on weekdays and invisible the rest of the time. When both moments are timestamped, the median is on the page before you scroll.
That is Sales Analytics: every lead's real timestamps turned into the numbers, with nobody typing a report.
Number two: how many days to a sendable quote?
Days from the site conversation to a priced document in the customer's hands. A 40x60x14 with a 12-foot lean-to is a day's takeoff at the outside; when turnaround runs eleven days the job is sitting behind three other takeoffs and a supplier callback. So watch the queue, not the speed. Five priced-but-unsent estimates is a five-day floor however fast anyone works, and the tell is a stage where the median sits far below the average — a handful of stuck deals dragging the whole stage.
Split the clock: days waiting on the customer's information, and days waiting on you. Only the second half is yours. If the takeoffs themselves are slow, how to bid a post-frame job without guessing is the read.
Number three: was the estimate opened?
The cheapest signal on the list and the most misused. A PDF emailed with no delivery signal cannot tell "they are thinking about it" from "it went to spam," and those call for different follow-ups: opened four times with no reply means they are working the number; never opened means fix the address before you write another word.
Open and view tracking on every estimate gives you the count and the seconds. Opens do not cause closes — buyers with real intent open more — so treat it as triage, never as a target.
Number four: what is the win rate by job shape?
An overall win rate is three businesses averaged into one figure. Illustrative arithmetic: 60 jobs priced, 21 won — 35%. Split it and you find 28 residential garages with 13 wins (46%), 22 ag buildings with 6 wins (27%), and 10 commercial jobs with an architect and 2 wins (20%). Price the losing cell in hours: ten commercial bids at six hours each is 60 hours to win two jobs — change how you bid it, what you charge, or stop taking it and move the hours to the cell closing at 46%. Where those leads come from is its own subject.
Number five: how many days from milestone to money?
Posts set, and a $14,000 draw comes due. Measure the days between the crew finishing and the money landing, then split it, because the halves have different owners.
| The clock | Who owns it | Typical cause when it stretches |
|---|---|---|
| Milestone done → invoice sent | Entirely you | Nobody told the office the posts were done |
| Invoice sent → paid | Shared | Terms, bank timing, a disputed line |
Measure the first half first: a crew finishing Thursday and an invoice going out the following Wednesday is six days of your own money financing the job. When the draws already exist as invoices off the agreed schedule, the unsent one is a visible draft beside the paid ones.
That is one invoice per draw off the schedule the customer approved. The other quiet source of unpaid work is scope built before it was papered — the extra overhead door, the slab thickened for a hoist — covered in the change order that protects your margin.
Number six: estimate against actual on which three lines?
Not whole-job gross margin; it arrives too late and too coarse. Track the post and lumber package, the steel package, and labour hours to dry-in — three numbers, three minutes at closeout. What this catches is unit mismatch: a supplier quoting per lineal foot while your sheet assumes per piece, steel priced per square while your takeoff counts panels. A unit error never shows up as one bad job; it shows up as the same slice missing from every job, and only the line-item comparison surfaces it. Job profit — estimate against actual keeps that comparison open while the job is still running, inside Project Command.
What do the shops with their act together do differently?
They measure the queue, not the person. They fix the denominator once. They read distributions, not averages. They act on one number a quarter — pick the worst, change one thing, leave the rest alone long enough to see whether it moved — and accept that sixty jobs a year is direction and outliers, not significance.
None of it tells you whether your price is right. These six measure the operation — how fast it responds, how cleanly it converts, how quickly it collects — and say nothing about your position in the market; you can be the tightest shop in the county and still leave money on every ag building because your rate has not moved in three years. Work the numbers you own, and treat pricing as its own argument every spring.
