There’s no published, industry-wide arbitration loss rate to hold your number against, the way NAAA publishes a hard $800 defect threshold. What actually matters is whether your loss rate is trending down against your own history and against real peer data from a 20 Group or your auction’s own composite report, not a figure pulled from a blog post.
That’s an unsatisfying answer if you’re the one who owns the number and just got asked in a QBR whether it’s “normal.” But it’s the honest one.
Failure mode
Chasing a mythical benchmark percentage is worse than not having one: it lets you feel fine about a bad trend or panic over a good one, depending on which stray number you compared yourself to.
Why there’s no single number to benchmark against
NAAA’s arbitration policy is public. Anyone can read the dollar thresholds, the claim windows, the light system, and the disclosure carve-outs, because the NAAA policy exists to be a shared rulebook that both buyers and sellers operate under. Arbitration outcomes, by contrast, are not public. Loss rate lives inside seller-facing portals at Manheim, ADESA, and ACV, inside a remarketing company’s own dashboards, and inside the composite reports that circulate through NIADA and NADA 20 Groups. Nobody aggregates that into a trade-press number the way, say, average days-to-sale sometimes gets reported, because arbitration loss rate is close to a direct reputational signal for the seller behind it, and sellers have no incentive to make that comparison easy for competitors or buyers to find.
That absence gets filled by folklore. Someone repeats a number they heard at a conference, or a vendor cites a round figure with no source, and it circulates as though it were a fact. Treat any specific “industry average” arbitration loss rate you see quoted without a named source (a 20 Group composite, your own facilitating auction’s aggregate reporting, a cited NIADA or NADA study) as unverified. It usually is.
What actually predicts whether your number is high
Since there’s no external benchmark worth trusting blindly, the more useful question is what drives the number up or down inside your own book, because that’s what you can actually act on.
Unit mix. A book that’s heavy on repo and trade-in inventory, with more accident history and title complexity per unit, will run a structurally higher arbitration exposure than an off-lease program with clean, single-owner records. Comparing your raw loss rate against a peer running a different mix isn’t a benchmark, it’s an apples-to-oranges number that will make one of you look worse than the underlying operation actually is.
Channel. Digital and physical lanes carry different risk profiles: a buyer inspecting a unit in person before bidding catches some issues that a photo-and-report listing can’t surface until after the sale. If your loss rate moved because your channel mix shifted, that’s a mix change, not a quality change, and the two need different fixes.
Root cause. This is the split that tells you where to actually spend effort. A loss rate driven mostly by condition-report mismatches (something the inspector missed or under-described) is a different problem than one driven by title and lien issues, which is different again from missed disclosures on required announcements. Our guide to improving condition report accuracy covers the first category in depth; most remarketing teams already track total loss rate but far fewer break it down this way, which is exactly the breakdown that turns a scoreboard number into a diagnosis. The prevention playbook walks through where each of these failure modes actually enters the pipeline, upstream of the sale itself.
Say your operation lists 300 units a month and eight come back as arbitration losses. That’s roughly 2.7%. Whether that’s a problem depends entirely on whether six of those eight are the same root cause repeating, or six different one-off misses across six different inspectors. The first is a fixable process gap. The second might just be noise in a small sample.
Key insight
A single monthly percentage collapses that distinction; a root-cause breakdown restores it.
How to get a real comparison
If you want an actual peer number, not a rumor, three sources get you closer than anything published externally:
-
Your own trend, over a rolling period. Arbitration loss rate is typically measured as a percentage of listed units resulting in a claim the seller loses, tracked over a rolling window (commonly a trailing 30, 60, or 90 days) rather than a single month, since monthly volume swings can make a small sample look noisy. Your own six-month trend line is more informative than any external number, because it holds your unit mix and channel roughly constant.
-
Your facilitating auction’s seller-level reporting. Manheim, ADESA, and ACV all surface some version of seller performance data through their seller portals. That’s the closest thing to an apples-to-apples comparison you’ll get, since it’s calculated the same way for every seller on that platform.
-
A structured peer group. NIADA 20 Groups put 20 non-competing dealers together specifically to exchange operational data and identify where each member stands relative to the group. NCM Associates, one of the larger 20 Group operators, runs custom monthly composites built from member-submitted data specifically so a dealer can see how their numbers compare against peers and against the wider group, not just against their own history. That’s a real benchmark, because it’s built from actual peer submissions under a consistent methodology, not a number someone half-remembers from a panel discussion.
| Source | What it actually tells you | Limitation |
|---|---|---|
| Your own rolling trend | Whether you’re improving or worsening against your own baseline | Says nothing about how you compare to peers |
| Auction seller portal | How you compare to other sellers on that specific platform | Only covers volume that moves through that auction |
| 20 Group composite (NIADA, NCM, similar) | A real peer comparison under a shared methodology | Only as good as your group’s mix similarity to your operation |
| An unsourced “industry average” | Nothing verifiable | Not a real number until you can trace where it came from |
None of these gets you a single universal “normal” percentage, because that number doesn’t exist. What they get you is a comparison that actually means something for your specific mix, channel, and volume.
Why this number feels personal, and why that’s not irrational
Arbitration loss rate carries more weight than most operational KPIs for a specific reason: it’s directly comparable across operations in a way most metrics aren’t, and it functions as a visible signal inside a buyer community that’s smaller and more networked than it looks from the outside. A high inventory-turn number is a mildly disappointing metric. A high arbitration loss rate reads closer to “you sold me a lemon and didn’t tell me,” and that’s not an abstract KPI miss, it’s reputational among repeat buyers who remember exactly which sellers make them inspect more carefully next time.
There’s also a real fairness problem underneath the anxiety, and it’s worth naming because it’s usually true: the person who owns this number rarely made the mistake that caused the loss. She inherited an inspection that happened upstream, sometimes at a different site, sometimes weeks before the unit ever listed, and gets measured on an outcome she didn’t personally create but is fully accountable for when it’s reported up. That’s exactly why the root-cause breakdown matters more here than in most KPI reviews: it’s the difference between explaining a systemic disclosure gap you’re actively fixing, and being unable to explain anything beyond “the number went up.”
If arbitration loss rate is climbing alongside recovery rate slipping or days-to-sale stretching, those three usually share a root cause worth investigating together rather than separately; our guide to recovery rate and days-to-sale as remarketing KPIs covers how they interact. And if the immediate question is what to actually do about a rising loss rate rather than how to interpret it, how to reduce auction arbitration claims is the more tactical companion piece to this one.
FAQ
How is arbitration loss rate typically measured?
As a percentage of listed units that result in an arbitration claim the seller loses, tracked over a rolling period (commonly trailing 30, 60, or 90 days rather than a single calendar month) and often benchmarked against peer operations through a 20 Group composite or an auction’s own seller-level reporting.
Why does this metric carry more weight than other KPIs?
Because it’s directly comparable across operations in a way most internal metrics aren’t, and because it functions as a visible, personal reputational signal inside a relatively small community of repeat buyers who track which sellers disclose accurately. A single bad number reads as a character judgment, not just an operational miss.
What’s a good arbitration loss rate?
There’s no published figure to answer that with, so the more useful question is whether your rolling trend is improving, and whether your root-cause breakdown is dominated by one fixable pattern or scattered across unrelated one-off misses. Compare against your own history first, then against a real peer source like a 20 Group composite, and treat any unsourced “industry average” as noise.
Decomposing arbitration losses by root cause, and catching the condition or title mismatch before a unit goes live rather than after a buyer files a claim, is the kind of pre-listing verification work Deskflow is built to run without adding headcount to the queue.