A leak looks small against revenue and large against profit.
Four leaks · two-minute estimate · your numbers, not ours
Revenue rarely leaves through one big hole. It seeps out through slow follow-up, quiet churn, soft pricing, and rework nobody logs. Each one looks like a rounding error against revenue, which is why each one gets waved off, and waved off correctly.
Then divide it by your net margin. Same money. Two denominators. One of them makes it invisible and the other makes it the largest single item on your desk.
These are starting points, not benchmarks. Nobody publishes a reliable leak rate for a business your size, and anyone who hands you one is selling you an average. Move each to what you think it is. The block below the results tells you how to measure it properly.
What the research measures
Construction. A study of 19,605 rework events across 346 projects, one contractor over seven years, found rework cost 0.39% of contract value. That same rework cut the contractor's mean annual profit by 28%. A later study by the same researcher found contractors underreport rework costs by 300%.
Love, Peter E.D. and others, "The Costs of Rework: Insights from Construction and Opportunities for Learning", 2018. Love, "Quantifying the Costs of Field Rework in Construction", Journal of Construction Engineering and Management Vol. 152 No. 1, American Society of Civil Engineers, October 2025.
Manufacturing. Cost of poor quality averages roughly 15% of the sales dollar, range 5 to 35% by product complexity. One action-research study found visible quality costs of 5.64% of sales and hidden costs of 8.78%, the hidden portion 1.6 times larger than what the accounting system captured, and the combined 14.42% exceeding the firm's entire net profit.
Institute of Industrial and Systems Engineers, summarising J.M. Juran, Quality Planning and Analysis, 3rd edition.
Professional services. Three rates multiply and nobody multiplies them. Utilisation near 37%, realisation near 88%, collection near 91%. Under thirty cents of every working hour reaches the bank.
Clio, Legal Trends Report, annual. Clio sells the practice-management software the firms in question bill with, and publishes the industry's most-used benchmark from anonymised platform data. We name that interest because it is what makes the number checkable.
Each figure above keeps the denominator its authors used. None has been converted into a share of your revenue, because that conversion would be ours rather than theirs.
What the research measures
No Canadian source measures this by sector at your revenue. Use your own last quarter: quoted value against invoiced value, on the jobs you won.
What the research measures
No Canadian source measures this by sector at your revenue. Use your own last quarter: enquiries received against quotes issued, and how many days sat between them.
What the research measures
No Canadian source measures this by sector at your revenue. Use your own last quarter: customers who bought last year and not this year, valued at what they used to spend.
Four industries, four names for it. Manufacturing calls it cost of poor quality. Construction calls it rework. Professional services calls it realisation. Software calls it churn. They are all measuring the same thing, the gap between what was sold and what was delivered, and every one of them books it somewhere the owner never looks.
Money that already left. Rework done and absorbed, discounts given, work delivered below the price it was quoted at.
Recovers near contribution margin where you have the capacity to do the work.
Work you could have done and did not. Enquiries answered late, quotes never chased, jobs that went elsewhere while yours sat.
Recovers near contribution margin only where capacity exists. If capacity has to be added, the rate falls toward gross margin and payback lengthens.
Still on your books. Revenue the same pattern takes next year if nothing changes.
Not yet lost, and not counted as recoverable. Shown separately because it is the only tier you can still stop.
Three totals, not one. Tier one has already gone. Tier two is capacity you are carrying and not filling. Tier three has not happened yet. Only the first two are money you can go and get, and only where you have the capacity to do the work.
Against revenue, all of this is a rounding error. Against profit, at a 7.0% net margin, it is not.
One assumption you should correct if it is wrong. These figures assume your crew is retained. Where labour is hired against jobs, the recovery rate is roughly twenty points lower. Which applies is your answer, not ours.
The number above is your guess. Here is how to replace it with a measurement.
It takes a week, and you do not need us for any of it.
- Pull your last four closed jobs.
- For each one, sit down with the person who ran it and ask a single question: what did we do on this job that never got written up? He will remember. People always remember. They can name the trip, the day, the material, usually to the hour.
- Put a rate against the hours and a cost against the material. Total the four.
- Divide by the contract value of those four jobs. That is your percentage, measured instead of guessed.
- Divide that by your net margin. That is the number that matters.
Two things usually show up. The guess was low, because the items were small one at a time, which is exactly why they were absorbed. And nobody was hiding anything. These are competent people making a sensible call each time, with no way of seeing the total they were building.
The net margin
Statistics Canada, Financial Performance Data, 2024 reference year, businesses filing $5,000,001 to $20,000,000, incorporated, Canada. Built from the GIFI schedules attached to T2 corporate returns, so it is a census of filed returns rather than a survey.
Four things worth knowing before you use it
Statistics Canada flags each line A through E for quality, and where a line we use is flagged E we say so. Each line is averaged only across businesses that reported it, so components will not sum to the total. Filed data stops at $20 million and no Canadian source replaces it above that line. And any EBITDA figure we show elsewhere is calculated by us from filed lines rather than published as one.
The leak percentages
Yours. Nobody publishes a reliable leak rate for a business your size. The research beside each slider measures related things in its own terms, and we have not converted any of it into a share of your revenue, because that conversion would be ours rather than theirs.
Questions We Hear About This Calculator
Where do these numbers come from?
The net margin is filed Canadian tax data: Statistics Canada Financial Performance Data, 2024, businesses at $5 million to $20 million, incorporated. The leak percentages are yours. Nobody publishes a reliable leak rate for a business your size, so the tool asks you for an estimate and then shows you how to measure it properly on your own last four jobs.
Why show the leak as a share of profit instead of revenue?
Because that is where it lands. A leak worth one percent of revenue costs one percent divided by your net margin of your profit. On a ten percent margin that is a tenth of your profit. On a four percent margin it is a quarter. Same leak, different business, very different year. Expressed against revenue it looks like a rounding error, which is exactly why it survives budget meetings.
My business is over $20 million. Does this still work?
The arithmetic works. The default margin does not. Filed Canadian data stops at $20 million, and no Canadian source publishes cost structure by revenue band above that line. You can confirm that in two minutes. Above $20 million the default shown is a floor, so enter your own margin.
Why three totals instead of one number?
Because they recover differently and one of them has not happened yet. Confirmed loss has already gone. Forfeited capacity is work you could have done, and it only recovers if you have the capacity to do it. At risk is still on your books and can still be stopped. A single blended figure hides all three distinctions, and it is the first thing a controller takes apart.
What is a Revenue Blueprint?
A working session that maps how your business runs today and what it would take for it to run without you. You leave with the map either way. Read what it covers and what it costs.
Bring the measured number, not the estimate.
Run the week of homework above first. Then book thirty minutes and we will read your four jobs against your filed margin and tell you which room the money is leaving from.
- What you leave with: your measured leak rate, expressed against your own profit, and the one room it is concentrated in.
- How long it takes: thirty minutes, and it ends at thirty minutes.
- What it is not: no deck, no proposal, no pricing conversation. If you want one afterwards you can ask for it.
- What we do first: we pull your sector's filed margin and cost structure before the call, so the time goes on your numbers rather than on ours.
- When not to book it: if you have not run the four jobs, there is nothing to read and the thirty minutes will be a conversation about averages. Do the homework, or leave this alone.