Most construction and industrial firms do not lose margin on the job. They lose it on the bid, before a shovel is in the ground, in a function nobody is watching closely because it looks like it is working. The win rate is fine. The estimator is busy. The proposals go out. And the money leaks anyway.
You do not need a consultant to find out whether this is happening to you. You need an afternoon and an honest answer to five questions. Each test below is something you can run yourself, today, with the data you already have. None of them require new software. All of them require you to stop assuming the function is fine because nobody has complained.
A leak you cannot see is still a leak. It just bills you quietly.
Run all five. Count your fails. At the end, we will tell you what the count means.
/ TEST 01The clock
Pull your last twenty-five leads. For each one, find two timestamps: when the lead arrived, and when a human actually replied. Not the auto-acknowledgement. The real first response from a person who could move the bid forward.
Now average the gap.
- Under four hours: you are competitive on response. Move on.
- Four to twenty-four hours: you are losing the bids that go to whoever called back first, and you are not counting them as losses because they never became bids.
- Over twenty-four hours: the lead is half lost before your estimator opens it. You are paying to generate demand and then handing it to your competitor's faster phone.
The reason this one matters more than people expect: the leads you respond to slowly are not randomly distributed. They are the busy weeks, which are also the weeks you most needed the work to land cleanly. The leak gets worse exactly when you can least afford it.
/ TEST 02The one person
Ask yourself a single question and answer it honestly: if your lead estimator took three weeks off, unplanned, starting Monday, what happens to your bid pipeline?
If the answer is "it stops," you do not have an estimating function. You have an estimator. The difference is everything. One is a system that survives a person leaving. The other is a single point of failure that you have been calling a department.
The test inside the test: of your last twenty-five bids, how many were written at full effort, with site visits done, vendor pricing confirmed, and margin reviewed? If the honest number is under twenty, your one person is rationing their attention across more bids than they can carry, and the bids that get the half-effort version are the ones quietly losing.
The vacation test
Picture your best estimator out for a month with no notice. If your honest reaction is a knot in your stomach rather than a name of who covers it, you have found leak 02. The fix is not hiring a second estimator. It is building the function so that bids do not live in one person's head and calendar.
/ TEST 03The loss you cannot explain
Walk to your estimator right now and ask them to pull up the last bid your firm lost. Then ask for the documented reason it was lost. Time it.
- Under three minutes: you have a working feedback loop. Rare. Keep it.
- Three to ten minutes: the reason exists somewhere but it is not captured anywhere a pattern could form. You are losing the same way repeatedly and you cannot see the repetition.
- Over ten minutes, or "I would have to ask around": there is no loop at all. Every loss is a one-off in your firm's memory, which means none of them teach you anything.
This is the most expensive leak because it compounds. A firm without a win/loss loop does not get better at bidding over time. It gets older at bidding. The estimator accumulates instinct, but instinct retires when the estimator does, and the firm starts over.
/ TEST 04The everything pile
Of the last twenty-five bids you submitted, how many did you have a real, better-than-even chance of winning when you started writing them?
Most firms cannot answer this because they do not decide whether to bid. They bid. The pipeline is an everything pile, and the cost of writing the bids you were never going to win is invisible because it shows up as a busy estimator and a reasonable-looking win rate, not as a line on the P&L.
Here is the math nobody runs: if writing a serious bid costs you a real day of your best estimator's time, and you write ten you had no business writing, that is two weeks a year your most valuable bid resource spent losing on purpose. The win rate looks fine because the denominator is padded with bids you should have declined.
/ TEST 05The margin you assumed
Take your last five won jobs. For each, compare the margin you bid to the margin you actually finished at. Then ask one question: who decided the final number that went out the door, and what discipline were they working against?
In most firms, the same person estimates the job, prices it, and negotiates the close. That is three jobs with three different incentives collapsed into one person under deadline pressure. Margin does not leak in a dramatic moment. It leaks in a series of small, reasonable-sounding concessions that nobody reviews because the person making them is also the person trusted to make them.
If you cannot reconstruct, after the fact, why the final price was what it was, you do not have pricing discipline. You have a pricing habit, and habits drift.
Margin rarely leaves in a flood. It leaves in concessions nobody logged.
/ THE SCOREWhat your count means
Tally the tests you failed.
The point of the count is not to alarm you. It is to tell you whether the problem is a leak you can patch on a Friday afternoon or a function you need to actually build. Those are different problems and they cost different amounts to ignore.
/ NEXT STEPIf you want the deeper version
These five tests are the self-serve version of the diagnostic we run at the start of every Operating Sprint. The full audit puts a defensible dollar figure on each leak and ranks them, so you are fixing the one that costs the most, not the one that annoys you the most. We wrote up exactly how that runs, hour by hour, in a separate note.
But if you ran the five tests and you already know what is leaking, you may not need us at all. The frameworks are free for a reason. Fix what you found. If the fix turns out to need a system you do not have the time or the authority to build, that is when an operator earns their keep.
The frameworks are free. The work is the offer.