Tires before engines.

A used car at 50,000 miles isn't "halfway worn out." It's halfway through a dozen different lifespans — some systems have already peaked their failure window, others are about to enter theirs.

That isn't shop wisdom. It's what the federal complaint record says when you actually sort it by the odometer reading owners wrote on the form.

NHTSA's complaint database lets owners include the mileage at which a failure happened. We pulled every problem cluster across the database where at least ten owners did that, and we computed the per-cluster failure-mileage distribution: 25th percentile, median, 75th percentile. Then we aggregated those distributions by NHTSA component category.

Here's what we found.

The headline

Across the 15 well-sampled NHTSA component categories, the median odometer at first complaint runs from about 29,615 miles (tires) to about 74,792 miles (seatbelts).

That's a 2.5× spread. Tires hit at a third of the mileage that seatbelt complaints do. Brakes are mostly behind you by 80,000 miles. Engine and suspension defects haven't fully shown up yet.

The numbers are aggregated across 308,767 owner-reported mileages spanning 7,630 problem clusters in the database. Every cluster in the analysis has at least ten reported mileages — the small ones got filtered out.

Three phases of failure

The categories cluster into three groups when you sort them by median failure mileage. Each phase has a different character.

1. Wear systems — failures cluster before 55,000 miles.

Tires, wheels, brakes, plus the human-interface electronics that get used hardest (cruise-control, visibility). These are the systems where complaints come from a mix of normal wear and early-life defects. A tire complaint at 30,000 miles is usually a tread-life issue or a sidewall problem the owner blames on the OEM. Brake complaints in the 50,000-mile band are pull, pulse, or premature scoring — not just pads at end of life. The pattern: the parts that get the most use complain first.

2. Mid-life systems — failures emerge 55,000 to 65,000 miles.

Powertrain, steering, electrical, fuel-system, body. This is where most of the mid-warranty-exit complaints land. Powertrain at 61,000 miles is transmissions losing their shift quality, mounts wearing, transfer cases starting to whine. Electrical at 63,000 is the modules and sensors that don't fail catastrophically but throw codes the dealer struggles to clear. Steering at 62,000 is racks, columns, electric-assist motors. The mechanical-electrical guts of the car start showing pattern problems right where most factory bumper-to-bumper warranties end.

3. Late-emerging defects — failures appear after 65,000 miles.

Lighting, airbags, engine, suspension, seatbelts. The systems where pattern defects need miles to surface. Engine complaints at 70,000 are the head-gaskets, the oil-consumption rings, the timing-chain stretch — failures that take heat cycles, not calendar time. Airbags at 68,700 is the inflator-degradation pattern. Suspension at 70,100 is control arms, struts, and bushings failing in patterns no one noticed when the car was new. This is the band where the genuinely defective — not the worn — starts to show.

The maintenance roadmap that falls out of this

If you're shopping a used car, the position of the odometer matters differently than the calendar age.

At 30,000 miles: tires are the first thing in the failure window. The original OEM rubber has reached an age where the worst of the OEM-spec problems are reported. Check tread, sidewall, and noise. Brake complaints are just starting to climb.

At 50,000 miles: brakes, cruise-control, visibility are at the median of their complaint window. Test cruise across hills, run the defrost and wipers, drive enough miles to feel the brakes work hot. Powertrain complaints are starting to surface.

At 65,000 miles: mid-life systems are past peak. Body, electrical, fuel-system. This is also the band right where most factory warranties end — not coincidentally. If you're buying here, the documented per-vehicle failure record matters more than at any other mileage band.

At 70,000+ miles: engine, suspension, airbag defects are entering their reporting window. Pattern-defect vehicles (Theta II, 6F35, EcoBoost water pump, Subaru EJ25, Honda 1.5T) tend to show their tell here. Match the vehicle's documented pattern against where the inspection should focus.

The 15 categories, in full

Sorted by median failure mileage, ascending.

The other story: how wide the window is

The median tells you when failures cluster. The interquartile range — Q3 minus Q1 — tells you how predictable that clustering is.

Some systems fail in a tight window. Tires have a Q1–Q3 spread of about 33,000 miles. Wheels are about 37,000. When these systems fail, they tend to fail in a predictable mileage band, and an owner buying at 40,000 miles knows what they're walking into.

Other systems fail across a much wider band. Seatbelts have a Q1–Q3 spread of nearly 59,000 miles. Airbags and suspension are both above 56,000. These are systems where the complaint mileage is essentially "anywhere from 40,000 to 100,000" — the failure isn't tied to one wear event, it's a long-tail pattern that can show up almost any time. Plan accordingly.

How we did it

Source. NHTSA's complaintsByVehicle public API. Synced weekly. We compute mileage statistics per problem cluster — the per-vehicle, per-component-category aggregation we generate from the raw complaint stream.

Per-cluster statistics. For each cluster, we record the 25th percentile, median, and 75th percentile of the reported odometer readings on its complaints. Mileage_sample_size is the count of complaints in the cluster that included a readable odometer.

Category aggregation. Filtered to clusters with at least 10 reported mileages (the floor for a stable per-cluster percentile). Then averaged each percentile across the qualifying clusters in the category. Result: 7,630 clusters, 308,767 mileage data points, distributed across 15 categories. (The 16th, "equipment", had only 4 qualifying clusters and 88 samples — we excluded it from the chart and main analysis but include it in the underlying dataset.)

The three-phase grouping is editorial. We sorted the 15 categories by median and split into roughly equal terciles. Other groupings would tell related stories.

What this doesn't tell you

This isn't failure rate. It's the mileage at which complaints are reported, weighted by what owners chose to file with NHTSA. A category with a low median failure mileage isn't necessarily failing more often than one with a high median; it's failing earlier in the cars where it fails.

Owner-reporting bias is real. Owners disproportionately file complaints about systems where they feel the manufacturer failed them — safety equipment, drivetrain, dealer-stonewalled repairs. Routine wear items get under-filed. The tire result is partly a real safety-recall signal (Firestone-era patterns, sidewall failures), partly an artifact of OEM-spec rubber being easier to blame than your driving.

The slim production database caps each vehicle at 20 complaints. For very-high-volume clusters, the mileage distribution we compute is from the most recent 20 complaints, not all of them. The full database has every complaint, and the percentiles change slightly when computed against it. For a piece at this aggregation level, the cluster medians are stable.

Mileage is self-reported. Some owners round. Some forget the exact number. The 25th and 75th percentiles are more robust to that than the mean would be, which is why we use them.

Reproducing this

The data is public. NHTSA's complaints API is free. Our per-cluster mileage statistics are computed in the sync pipeline and exposed on each problem page — the histograms and decay curves on individual problem pages show the per-cluster distribution that aggregates into the per-category result here.

For the platform-level analogue — failure mileage by engine and transmission family rather than by NHTSA category — see When Things Break, in Actual Miles.