A single “win rate” number hides more than it shows. Count-based and value-based win rate can point in opposite directions on the exact same quarter of bidding, and neither one means much without a defined denominator.
Key takeaways
“What’s our win rate” sounds like a single-number question. Run it two honest ways on the same set of bids and it is often two different answers, each true, each telling a different part of the story — and reported as one blended figure, it can quietly hide the more useful of the two.
Win rate by count is the simplest version: bids won divided by bids submitted, expressed as a percentage. Win rate by value replaces both numbers with dollars: total contract value won divided by total contract value bid. The two will match only by coincidence — a firm can win a quarter of its bids by count while winning well over a third of the dollar value it chased, or the reverse, and the gap itself is the useful part of the measurement, not a rounding difference to be ignored.
No Statistics Canada release, no CCDC document, and no industry association this page could find publishes an average or benchmark win rate for Canadian contractors, by sector, by region, or otherwise — a check of Statistics Canada's own construction statistics library turns up building permits, capital expenditures and housing starts, and nothing resembling a bid-success metric. That absence matters: a win rate only means something compared against a firm’s own trend over time, not against an invented “industry average” that does not exist in any citable Canadian source. Treat any number presented as a national benchmark with the same scrutiny this page is applying to itself by refusing to invent one.
“Bids” has to mean one specific thing before a win rate means anything at all: invited-to-bid opportunities, bids actually submitted, or every opportunity a firm looked at including ones it declined to pursue. Declined opportunities belong in a separate go/no-go tracking measure, not in the win-rate denominator itself — folding them in quietly lowers the rate without telling anyone why, and makes quarter-over-quarter comparisons meaningless the moment the firm’s appetite for bidding changes.
A blended win rate across every sector and procurement route a firm bids into smooths over exactly the pattern worth finding. Splitting results by sector (institutional versus private, say), by procurement route (stipulated-price tender versus design-build RFP), or by size band usually shows one segment performing well above the blended average and another dragging it down — information a single number simply cannot carry. A firm that has already built an estimating system that scales across bid volume has the underlying job-level data to segment this way without extra data entry; a firm still tracking bids in an ad hoc way does not.
Total addressable bid volume moves independently of any one firm’s skill at bidding — Statistics Canada’s regional and city-level construction data (see building permit values by province) is one real, citable way to see that the market itself expands and contracts. A falling win rate in a quarter where the whole regional market contracted is a different finding than a falling win rate in a stable or growing market, and conflating the two — blaming the estimating team for a market-wide slowdown, or crediting them for a rising tide — is a common and avoidable misread. Checking the regional data before the internal debrief even starts is a cheap first step that either rules out or confirms a market-wide explanation before anyone spends time reviewing individual bid files for a cause that was never there.
Twenty bids in a quarter is not a large sample, and a firm bidding fewer than that — common for firms working larger, less frequent contracts — should treat any single quarter’s win rate as noisy rather than diagnostic. One large win or one unusual loss can move a count-based rate by five or ten points on a small sample without reflecting any real change in competitiveness. A rolling trailing-four-quarter view smooths that noise out far better than reacting to whichever single quarter just closed, and is usually a more honest basis for a resourcing or estimating decision than a single period taken in isolation.
A win rate answers how often; it does not answer why. The two measurements are complementary rather than redundant — a structured post-bid debrief is what turns a falling win rate into a specific, named cause (a compliance gate, a technical scoring gap, or genuine price competitiveness), and a win-rate trend is what tells a firm whether the fixes coming out of its debriefs are actually changing results over time, rather than producing action items that sound right in the room and then quietly go nowhere.
A hypothetical contractor submits 20 bids in a quarter and wins 5 — figures constructed to illustrate the method, not a real firm’s results. Count-based win rate: 5 ÷ 20 = 25%. The total value of all 20 bids submitted was $18,400,000; the value of the 5 won was $6,900,000. Value-based win rate: $6,900,000 ÷ $18,400,000 = 37.5%.
The two numbers diverge because the firm is winning disproportionately larger jobs: average bid size across all 20 submissions was $18,400,000 ÷ 20 = $920,000, while average size of the 5 wins was $6,900,000 ÷ 5 = $1,380,000 — 50% larger than the typical bid. Reported as a single blended figure, either the 25% or the 37.5% understates the real pattern: this firm is not winning a quarter of its opportunities at random, it is winning disproportionately on its larger bids and losing more often on its smaller ones, which points toward a very different resourcing question (is the estimating team spending too little time on small jobs, or is the firm simply not competitive at that size) than a single flat win-rate number would ever surface.
Pick one definition and hold it constant quarter over quarter; submitted bids is the cleanest denominator, since it excludes both opportunities never pursued and pure expressions of interest that never became a priced number.
No — declined opportunities belong in a separate go/no-go tracking measure (see whether a firm has to bid every invitation it receives), not folded into the win-rate denominator, where they would silently distort the rate without explanation.
Quarterly is a reasonable operating cadence for most firms' bid volumes — frequently enough to catch a developing pattern, infrequently enough that a single unusual bid doesn't swing the number — but firms bidding fewer than roughly 15 to 20 jobs a quarter should read any single quarter's figure alongside a rolling trailing-four-quarter view rather than in isolation; this is a practical operating suggestion, not a figure drawn from any published standard.
A 30-minute call is enough to tell you whether your losses cluster by size, sector, or route.
