Your days supply is measuring cars you don't have.

The short answer: days supply is a two-number ratio, and almost every store gets the numerator wrong. Sold units that were never cleared, duplicates arriving from a second feed, cars sitting in recon that nobody status-changed, wholesaled units still showing "Available" — they all count as inventory in the report, and none of them can be sold to a retail customer tomorrow. You end up planning against a number that describes a lot you don't have.

This is not a repricing argument. Nothing here is about what to ask for a car. It's about the one thing every pricing decision, every buying decision and every aging report silently assumes: that the list of cars is correct.

What days supply actually measures

Days supply answers one question: at the rate I'm currently selling, how long would it take to sell everything on the lot if I never bought another car?

Most stores run it on a trailing 30 or 45 days. Use whichever you like — just use the same one every month, because a days-supply figure only means something against its own history.

Worked, with round numbers:

• 120 used units in stock
• 90 used units retailed in the trailing 30 days
• 120 ÷ 90 = 1.33 → × 30 = 40.0 days supply

Forty days. Comfortable, by most standards. Now hold that number, because we're going to break it.

What a good days supply looks like — and why the target moved

The classic operator rule of thumb has been 45 to 60 days on used. That band is older than the market it describes. Used days' supply has been widely reported running under 40 days nationally through 2026, and stores that turn fastest often deliberately sit lower still — 30 to 35 days, buying constantly to hold it there.

Treat all of those as commonly-cited reference points, not law. Days supply is market-specific and franchise-specific: a rural truck store and an urban compact store with the same 40-day figure are in completely different positions. The number that matters is your number, measured the same way, moving in a direction you chose.

What is universal is the direction of the cost. Every extra day of supply is a day of floorplan interest and a day of depreciation on every unit carrying it. NCM Associates' 20 Group work puts total holding cost near $37 per vehicle per day, and most dealers land somewhere in the $30–$40 range once floorplan, depreciation and opportunity cost are counted. You can estimate what the aging is costing you directly.

The four ghosts on every lot list

Here's the part nobody audits. A car can be on the inventory list and not be on the lot. Four ways it happens, all mundane:

• Sold but never cleared. The deal was written Saturday; the unit's status still says Available on Wednesday because the paperwork moved before the system did. It keeps aging on the report the whole time.
• Duplicated across feeds. The same VIN arrives from two sources with the stock number formatted differently — one system writes it in caps, the other in title case, one pads it, one doesn't — and your report groups them as two cars. One physical unit, two lines, two sets of days-in-stock.
• Never actually available. Units in recon, units awaiting a safety, units on a transporter, units with no clean title yet. Real cars, genuinely yours, but a retail customer cannot buy one tomorrow — which is the exact question days supply claims to answer.
• Wholesaled or deleted upstream and still listed. The unit left the system that owns it, but the downstream report never got the deletion, so it lingers indefinitely. These are the worst ones, because they never age out — they just accumulate.

None of these is anybody's fault. Every one of them is a status that was true when it was written and stopped being true later.

What the ghosts do to the arithmetic

Back to the 120 units. Say an audit finds 14 of them are ghosts — three sold-not-cleared, five duplicates from the second feed, four sitting in recon, two wholesaled last month and never removed. Real, retailable stock is 106.

Read what that actually costs you. You are carrying 14 fewer retailable units than your own report says — about 12% short. Forty days says "we're fine, slow down at the auction." Thirty-five says "keep buying." One of those is a decision to stop feeding the lot in the exact week you should be feeding it hardest — and the effect compounds, because under-buying today is next month's aging problem in reverse.

Then look at the aged line, which is worse. Ghosts skew old by their nature: a sold unit that was never cleared keeps accruing days forever, and a duplicate carries the same age as its twin. So they cluster in the aged bucket and inflate it — 20% of the lot over 60 days when the true figure is 14%. Your must-move list is padded with cars that don't exist. That's not just a wrong percentage; it's merchandising spend, price reductions and sales-meeting attention aimed at VINs nobody can sell, while the genuinely aged units get a smaller share of the room.

When two systems disagree, it's usually a definition

Every dealer group has had this meeting: the DMS says 441, the reporting tool says 472, and everyone assumes something is broken.

It usually isn't. The gap is almost always a status-definition difference — one system counts units in draft, or units on an expired sale, or in-transit, or consignment, or a third-party feed that isn't your rooftop at all, and the other doesn't. Nobody's data is wrong; the two systems were asked different questions.

The fix is boring and permanent: decide once, in writing, which statuses count as lot-ready, apply that definition everywhere, and reconcile to it — then a count difference genuinely means something is broken, and you can act on it. This is the same discipline the multi-rooftop reporting problem needs on the sales side, where "a sale" means booked in one store and delivered in another.

Reconciliation, not another report

You cannot fix this by looking harder at the report. The list has to be reconciled against the systems that own the truth, on a schedule, without anyone remembering to do it:

• Sync the feed, don't re-key it. Every unit's status comes from the system of record, nightly, not from someone updating two places.
• De-duplicate on the VIN, not on the stock number. Stock numbers get formatted differently by every system that touches them. The VIN doesn't. Normalize case and padding before you group anything, or you will count the same car twice forever.
• Honour deletions. If a unit disappears from the upstream feed, it needs to be pruned downstream — with a guard so that a failed or empty sync never mass-deletes a lot. A deletion you don't process is a ghost you keep forever.
• Separate "in stock" from "retailable." Recon, in-transit and untitled units belong in the count of what you own, and out of the count you divide by. Report both; never confuse them.
• Age from a date you trust. Days-in-stock has to run off the date the unit actually landed, not a field that gets blanked, back-dated or reused when a stock number is recycled.
• Put accuracy on the board. Inventory accuracy — real units versus listed units — belongs on the daily number set with everything else, or it will only ever get looked at during a physical count. It's number 10 on the twelve KPIs a GM should see every morning.

The same failure runs through the other tables in your business, for the same reason — a field that was true when someone typed it and quietly stopped being true. On the lead side it shows up as a lead source that says "Walk-In" when it wasn't. Different table, identical disease.

What good looks like

Whatever you use, require all six:

• One lot list, fed from the system of record, not maintained by hand
• Duplicates collapsed on VIN, with stock-number formatting normalized before grouping
• Upstream deletions processed automatically, with a guard against empty-sync wipes
• "In stock" and "retailable" reported as two separate numbers
• Days-in-stock aging computed from a date that can't be reused or blanked
• Days supply and aged-% recomputed off the reconciled list — not off the raw feed

Do that and the aged report becomes a work list instead of an argument. And the number you take to the auction is the number you actually have.

FAQ

How do you calculate days supply of used cars? Divide the units in stock by the units you retailed in a trailing period, then multiply by the number of days in that period. With 120 units in stock and 90 retailed in the last 30 days: 120 ÷ 90 × 30 = 40 days supply. Use the same trailing window every month so the figure is comparable, and divide only by units that are genuinely retailable — sold-but-uncleared units, duplicates and cars still in recon inflate the numerator and make the number look healthier than the lot really is.

What is a good days supply for a used car lot? The long-standing operator rule of thumb is 45 to 60 days, though used days' supply has widely been reported running under 40 days nationally through 2026, and fast-turning stores often hold 30 to 35 deliberately. Treat those as reference points rather than targets — the right figure depends on your market, your franchise and how quickly you can replace what you sell. What matters more than the benchmark is measuring it the same way every month and knowing the underlying stock count is accurate.

What is a phantom unit in dealership inventory? A phantom unit is a vehicle that appears on your inventory list but cannot actually be sold — a unit sold and never cleared, the same VIN duplicated from a second feed, a car sitting in recon or in transit, or a wholesaled unit that was removed upstream but never pruned downstream. Phantom units inflate days supply, so the report tells you to stop buying when you should keep buying, and because they skew old they also inflate the percentage of inventory aged over 60 days, padding the must-move list with cars that aren't there.