What Canadian cars asked, and how long they sat
Pick a slice of the Canadian market and this page answers two questions about it. Did it sell — the share of vehicles gone within 14, 30 and 60 days, by how far the day-one ask sat from the market price for that exact vehicle, with the cars still sitting on the lot counted in. And how fast did the ones that left go — every completed listing as one square, asking price against days on the market. New and Used are never blended.
Choose a slice
Price this car
Every vehicle in this slice that arrived 30 to 60 days ago and that we watched arrive — including the ones still for sale, counted as not sold. Drag a price and the odds move with it.
What moving the price is worth
Across Canada, a used vehicle asking more than 10% under the market price for its own make, model, year and odometer band was gone within 30 days 47.7% of the time. One asking more than 10% over was gone 35.7% of the time — 12.0 points of difference, and it falls at every one of the six steps between them. The same cars, measured by median days among the ones that sold, look flat: 18.5 days against 21.3. That statistic only counts cars that left, so the entire cost of asking too much — the car still sitting on the lot — is excluded from it by construction.
New vehicles are genuinely flat, and that is not a limitation of the method: 31.1% against 28.7% at 30 days across the whole price range, on 152,426 vehicles. MSRP anchors a new car's price, so how far the ask sits from the market carries almost no information. The page measures this per slice rather than assuming it, and says so when it finds it.
| Used | New | |||||
|---|---|---|---|---|---|---|
| Day-one ask vs market | Vehicles | Gone in 14 d | 30 d | 60 d | Vehicles | Gone in 30 d |
| More than 10% under market | 12,428 | 26.3% | 47.7% | 57.0% | 24,692 | 31.1% |
| 5–10% under market | 9,229 | 22.6% | 44.9% | 55.4% | 21,492 | 27.9% |
| 0–5% under market | 11,413 | 21.4% | 43.2% | 53.6% | 29,239 | 29.0% |
| 0–5% over market | 12,523 | 19.7% | 41.5% | 52.7% | 29,060 | 28.8% |
| 5–10% over market | 7,554 | 18.5% | 39.3% | 51.1% | 18,756 | 30.8% |
| More than 10% over market | 14,630 | 16.6% | 35.7% | 46.6% | 29,187 | 28.7% |
293,410 vehicles whose listing started between 09 July 2026 and 08 August 2026 — every one of them watched from the day it arrived, and 158,925 of them still for sale and counted here as not sold. Priced on the day-one ask against the median for the same make, model, year, condition and odometer band. The 14- and 30-day figures are complete for every vehicle; the 60-day one is a floor, because a car that arrived 35 days ago has had 35 days, not 60.
The evidence: what the cars that left actually did
One square per vehicle in the same slice, for the vehicles whose listing is over — asking price up the side, days on the market across the bottom. This is the half of the outcome the curve above does not need to be told about, and it is where a slice's spread, its outliers and its thin patches are visible one car at a time.
Asking price against days on market
Choose a make and the cloud draws itself.
What the four definitions change. About a third of these vehicles ever changed price at all — 34.0% of the new ones and 35.4% of the used, by an average of $4,100 and $2,567. So the middle barely moves: the median used car asks $28,981 on its last price, $28,995 averaged and $29,168 on day one — 0.6% across the whole range, against 0.2% for new. What moves is the tail and the spread, not the centre, and a median that shifts by a few hundred dollars here is a change of definition rather than a market that repriced. Time-weighted lands at $28,999 used and $52,434 new — exact for the 102,880 of 113,729 repriced vehicles whose change we can date (90%), and the plain average for the rest. The sell-through curve above is not affected by this control — it is always priced off the day-one ask.
New
Used
And what the ask did while they sat
Every listing is a two-step line: it asks its opening price until the day the price changed, then its final price until it goes. Stack the slice up and take the median at each day and you get the picture below — how much these cars actually come down, and when.
What the ask did while the car sat
Median asking price at each day of a listing's life.
Three things this chart cannot do, said here rather than left to be discovered. The date we hold is the last price change, so a car that cut three times is drawn as one step: every line is a lower bound on how much repricing went on, right about the endpoints and wrong about the number of moves. (api.market_history.price_event, the real change log, closes that gap — it is two days old and still mostly seed rows, so nothing here reads it yet.) The 60-day window bounds all of it, and 22% of the inventory currently for sale has already been listed longer than that, so the slowest cars on the market are not in this picture at all. And a price change we cannot place inside the listing's own life is dropped rather than nudged into range — the count is in the line under the chart, never folded away.
Why the pooled line is the one to be careful with. At day 40 the only cars left are the slow ones, and a slow car is likelier to have cut its price — so a pooled line that falls is falling partly because the vehicles under it changed, not because anybody repriced. That is the same trap as reading days-on-market off the cars that sold, which this page fixed once already. The at-risk row under the chart is there so the thinning is visible instead of implied; “by how fast they left” is the version that answers it, because a cohort is a closed group and its line is drawn solid for as long as not one vehicle in it has gone.
How to read it
A square low and to the left is a car that was priced to move and moved. High and to the right is a car that sat at its number. The shape worth looking for is the vertical spread at a given number of days — where the same model leaves in the same fortnight across a $15,000 band, price was not what decided it.
The two crosshairs are the medians, one per condition, labelled on the chart. Hovering any square gives you that exact vehicle: year, trim, odometer, province and the day it came off the market. Everything the hover shows is in the table under the chart as well, so none of it is locked behind a pointer.
New and Used are separate colours because they are separate markets, and a median across both describes no car anyone can buy: a blended Alberta F-150 in this corpus reads $75,650 at 25 km, which is neither the New figure ($79,385) nor the Used one ($47,990 at 87,730 km).
Why the cloud looks flat
The obvious thing to look for in the scatter is a slope — cheap cars going quickly, dear ones sitting. It is not there, and that is a finding rather than a fault in the chart. Holding year, model, trim, condition and odometer band constant, the cheaper half of a cohort took a median 22.6 days to go and the dearer half 21.8. Comparing vehicles that left against vehicles still unsold after two months, their average price rank inside their own cohort is 0.49 against 0.47. Among the cars that sold, asking price is not what decided how long they took.
That sentence has to keep its qualifier, because the qualifier is the whole story. "Among the cars that sold" throws away every car that did not sell, and that is exactly where the cost of asking too much lives. Count those cars — as the curve at the top of this page does — and the same vehicles over the same window say something quite different: a used car asking well under market is gone within 30 days about half again as often as one asking well over. Two statistics, one population, and only one of them can see the lot at the end of the month.
There is a real limit underneath that, and it is worth stating plainly. This page can only show a vehicle that both arrived and left inside the window. 34% of the 645,515 vehicles for sale in Canada today (217,900 of them) have already been listed longer than the whole 60-day window, and the median car currently for sale has been up 46.1 days against a median of 18.1 days for the used ones that left. The genuinely slow half of the market is therefore missing from every slice, and if a price-versus-time relationship lives out in that tail, this chart is the wrong instrument for finding it. What it does show honestly is the shape of the half that moves.
The one thing that does track time on market is a price cut, and it runs the opposite way to intuition: vehicles discounted before they left took 61% longer to go than those never discounted. Cutting is what slow cars do, and it does not rescue them. That figure is on the Omnia Index, computed over the same population.
Across the whole window, used inventory clears 5.8 days faster: Used at a median 18.1 days and $28,981 asking, New at 23.9 days and $52,373. 29.9% of used vehicles were discounted before they left, against 18.4% of new ones.
By make
The sixteen busiest brands in the window, New and Used reported separately. Medians only — each of these is thousands of vehicles, and the chart above is where the distribution behind them lives.
| Used | New | |||||
|---|---|---|---|---|---|---|
| Make | Vehicles | Days | Median ask | Vehicles | Days | Median ask |
| Ford | 15,393 | 18 | $32,655 | 26,369 | 26.4 | $70,020 |
| Honda | 13,341 | 16.9 | $26,490 | 15,372 | 21.4 | $46,910 |
| Chevrolet | 10,036 | 19 | $30,388 | 18,569 | 25.7 | $51,819 |
| Toyota | 15,158 | 16.5 | $33,388 | 12,037 | 14 | $53,483 |
| Hyundai | 11,720 | 18 | $21,788 | 13,741 | 22 | $42,370 |
| Nissan | 10,016 | 18.9 | $21,795 | 8,090 | 25.9 | $37,797 |
| GMC | 6,475 | 19.9 | $42,392 | 11,514 | 28.2 | $81,079 |
| Kia | 7,435 | 18.1 | $22,400 | 10,173 | 19.4 | $42,730 |
| Mazda | 6,733 | 16.4 | $26,978 | 8,373 | 23.2 | $42,290 |
| Volkswagen | 6,797 | 18.2 | $25,888 | 7,273 | 27.6 | $47,295 |
| Ram | 5,510 | 20 | $38,995 | 5,275 | 33.1 | $84,375 |
| Subaru | 5,042 | 16.3 | $27,999 | 5,454 | 16.6 | $42,606 |
| Jeep | 5,852 | 21.4 | $29,884 | 4,403 | 33.9 | $57,870 |
| Audi | 4,269 | 20.2 | $37,977 | 3,807 | 37.1 | $73,490 |
| Mitsubishi | 2,726 | 19.7 | $23,998 | 4,735 | 29.5 | $39,998 |
| BMW | 3,599 | 18.2 | $42,467 | 2,166 | 26.4 | $74,406 |
Vehicles, median days on market and median asking price, for spells that both began and ended inside the observation window. The ask here is the average of the first and the last one, which is what the chart above opens on; its Price shown control swaps in the other three definitions.
What this is built from
452,040 vehicles across 133 makes, 1,838 models and 4,794 dealers — franchise and independent lots together, 31,382 of them independent — delisted between 11 July 2026 and 07 September 2026. Rebuilt every morning at 08:00 Mountain.
Delisted vehicles only. Days on market is only a number once the listing is over; an active listing's age is a floor that grows while you look at it, and the two do not belong on one axis. Cross-listed duplicates are removed — the same car on three rooftops of one dealer group is one vehicle, counted once from the original seller's listing, and a car moving between those rooftops is not a sale. By default the chart shows only spells that both began and ended inside the window, so neither endpoint is an artifact of when we started looking; the wider set is one checkbox away and is labelled where it matters.
Quality exclusions match the live listings feed clause for clause — a Used car with no odometer, a New car under $5,000, a missing make or model, a dealer we no longer crawl. On top of that, each slice you plot has its own price fence: vehicles more than three interquartile ranges from that slice's middle, measured in log space because asking prices are log-normal, are excluded and counted underneath the chart rather than quietly dropped.
The sell-through curve is a different population from the scatter, on purpose. It is every vehicle whose listing started between 09 July 2026 and 08 August 2026 and whose arrival we watched — 293,410 of them, 158,925 still for sale and counted as not sold — including 1,723 whose listing ended at one rooftop while the same car stayed for sale at another, which is a move and not a sale. Each one has therefore had a full 30-day run that we watched all of, which is what makes "did it sell" a fair question to ask of it. Every vehicle is scored against the median day-one ask for its own make, model, year, condition and odometer band, over cohorts of at least 20 vehicles; 73,207 sit in cohorts too thin for that and carry no ratio. Rebuilt every morning at 08:00 Mountain with the rest of this page.
The curve refuses to draw rather than draw a shape it cannot support. Under 400 vehicles in a condition, or under 40 in any one of the six price bands, it says so and shows nothing. Above that it still names no direction unless the gap between the cheapest and the dearest band is at least 8 points of 30-day sell-through and the six bands run one way the whole way. Those two tests were set by measuring them against new cars, where the answer is already known to be "no relationship": they leave 1 make in 27 claiming an effect on the new side, against 13 in 23 on the used side.
These are asking prices and time on market, not transaction prices. We measure what sellers ask and how long they ask it for. Nothing here is an appraisal or a valuation for lending, insurance or legal purposes — see terms. The full method, including what it cannot do, is published in the Omnia Index.
One inconsistency worth stating plainly: the medians on /desk, /price and /research are computed from the crawl ledger, which carries a narrower delisted population than the dealer records this page reads. Figures here can therefore differ from the same slice there. Reconciling them moves published numbers and is being done separately.
Updated 07 September 2026.