Symptoms of a Bad Starter: What 1,483 Owner Reports Describe

Every article about a failing starter prints the same handful of symptoms, and none of them is measured against anything. The federal complaint file is the one public record that carries an owner’s own account of a fault, at scale, with a date and a component tag attached. So we pulled it: 165,858 owner complaints across 30 model lines and model years 2010 to 2020, and reduced them to the 1,483 whose narrative names a starter. The question this page answers is not “what are the symptoms” — anyone can assert those — but which of them actually distinguishes a starter from every other electrical fault, and what owners kept getting wrong on the way to finding it.

The loudest symptom is the least useful

A no-crank — you turn the key and the engine does not turn over — is the symptom that finally makes someone file, and it dominates the starter reports at 46.5%. But it is the least diagnostic thing in the file, because it runs at 11.6% in the control set too: a dead battery, a bad ignition switch, a security-system fault, a blown fusible link and a seized starter all end the same way, with an engine that will not crank. So the no-crank is the start of the diagnosis, not the answer. The 46.5% figure tells you owners reach for the government form when the car finally strands them; it does not tell you the starter did it.

The two noises that actually point at the starter

What separates a starter from the rest of the electrical system is sound. A grinding noise appears in 6.6% of starter reports and just 0.5% of the control — thirteen times as common — and it has a specific mechanical meaning: the starter’s pinion gear is clashing with the engine’s flywheel ring gear instead of meshing cleanly. A clicking noise runs 9.2% against 1.6%, nearly six times the control; a single heavy click is usually the solenoid throwing but the motor not turning, and rapid clicking is more often a supply problem — which is one of the ways the battery enters the story below. These two noises are the closest the file comes to a fingerprint. If you can hear a grind or a click when you turn the key, you have far better evidence than any no-crank alone provides.

The early warnings, and the symptom that isn’t one

Before the final no-crank, two patterns show up more often in starter reports than in the control and read as the failure warming up. The car started only after several tries in 29.4% of starter reports against 15.6% of the control, and the failure was intermittent — sometimes fine, sometimes not — in 21.3% against 14.6%. A starter that is on its way out often works when it is cool and refuses when it is hot, or works four times out of five, and those “it’s been doing it for weeks” accounts are the window in which it is cheapest to deal with.

One widely-repeated symptom is not one. “No warning — it just happened” appears in 12.5% of starter reports and 12% of the control: statistically level, so it is not a starter trait at all, it is simply how sudden electrical failures feel. Treat “no warning” as a description of the fright, not a clue to the part.

The battery gets blamed first — and replaced

This is the single most useful finding in the file, and it is about human behaviour, not the motor. A dead battery and a dead starter produce the same sentence — the car will not start — so the battery is what gets suspected. The battery is named in 38.8% of starter reports, twice the 19.8% rate in the control, and in 17.2% of them — roughly one in six — the owner describes a battery that was actually replaced. Only 2.2% record a battery that was tested good or explicitly ruled out before the starter was reached. In other words, a great many people bought a battery on the way to discovering it was the starter.

The practical reading is a rule you can use: a new battery that does not fix a no-crank, or a no-crank that comes back soon after a battery was replaced, is the signal to stop suspecting the battery and look at the starter and its supply. Have the battery tested before it is replaced, not after — the difference between our guides to the alternator, the battery and the starter is mostly which of them a proper test rules out first.

Why you cannot find it by component

A starter is electrical, but NHTSA’s filing does not treat it that way consistently. Among the starter reports, 62.4% carry the ELECTRICAL SYSTEM label — against 16.8% of the whole archive — but 38.6% are filed under ENGINE, and the rest scatter across power train, “unknown or other” and more. That is why this page is built from a text search of the narratives rather than from the component tag: filtering on ELECTRICAL SYSTEM alone would have thrown out nearly four in ten of the reports and produced a confident, wrong count. It is also a caution for your own research — a starter problem can be logged under half a dozen headings.

It is not a winter failure

The folklore says starters die on the first cold morning. The file does not agree. Starter reports fall 22.2% in winter (December to February) and 28.5% in summer (June to August) — and the control set runs almost the same, 24.2% winter against 27.8% summer. A marginal starter can certainly give up in the cold, but as a population these failures are no more a winter event than electrical trouble generally is. Do not let a summer no-crank talk you out of suspecting the starter.

What to do when you suspect it

Work the cheap, decisive checks in order. Have the battery load-tested first — not replaced, tested — because the file shows how often it is the innocent party that gets swapped anyway. If the battery is good and the engine still will not crank, listen: a single click or a grind when you turn the key points hard at the starter or its solenoid, while dash lights that come on normally but no crank at all narrows it further. Then price the repair with our guide to starter replacement cost before you authorise it, and if the diagnosis was a dead battery after all, our battery replacement cost guide covers that. For the wider no-start decision tree — fuel, spark, security and the rest — see why a car won’t start.

On a used car you are about to buy

A starter on its way out is easy to hide for the length of a test drive and expensive to inherit. Start the car yourself from cold, and listen for the click or grind rather than trusting a car a seller has left running. Because a starter fault can be filed under several components, it is worth checking the model’s record in the complaint lookup and its open recalls before you buy — a pattern of no-crank reports on that model year is exactly the kind of thing the federal file is good for, and exactly the kind of thing a seller will not mention.

Common questions

What are the symptoms of a bad starter?

In 1,483 owner reports, the commonest is an engine that will not crank (46.5%), but the symptoms that actually distinguish a starter are the noises: a grinding sound (thirteen times as common as in other electrical complaints) and a click (nearly six times). A car that starts only after several tries, or intermittently, is the early warning. “No warning” is not a starter-specific sign — it runs at the same rate across all electrical faults.

How do I know if it’s the battery or the starter?

Have the battery load-tested before replacing it. In the complaint file the battery is named in 38.8% of starter reports and was actually replaced in 17.2% — about one in six — while only 2.2% record it being tested good first. So a new battery that doesn’t fix the no-crank, or a no-crank that returns soon after a battery change, points past the battery to the starter. A single click or a grind when you turn the key points at the starter directly.

Can you drive with a bad starter?

Only until it fails to start next time, which the file suggests is unpredictable: 29.4% of reports describe a car that started only after several tries and 21.3% an intermittent failure, meaning it works until it doesn’t. Once a starter is clicking or grinding it can strand you without notice — 22% of these reports end in a tow — so treat it as a soon-not-someday repair.

Do starters fail more in cold weather?

Not according to the complaints. Starter reports lean slightly to summer (28.5%) over winter (22.2%), and the control set of other electrical complaints runs almost identically — so as a population these are not a cold-weather event. A marginal starter can still fail on a cold morning, but a summer no-crank is no reason to rule the starter out.