Think about the last five-star review you wrote for a mechanic. Now ask what it was actually praising. Odds are it said something like "Super friendly, got me in fast, explained everything, fair price!" Look at that sentence again. Not one word of it can be checked. You don't know whether the brakes needed replacing. You don't know whether the part was installed. You trusted a kind person who spoke clearly, and you wrote down the feeling.

That isn't a flaw in you. It's the central problem with reading mechanic reviews, and once you see it, you'll read every star rating differently.

The product you can't grade

In 1973, economists Michael Darby and Edi Karni described a category they called credence goods. These are services where the seller knows far more than the buyer, and where the buyer often can't tell whether the work was needed or done properly even after it's finished.

Car repair is the textbook case. When your car stops making the noise, you feel relief. But did the shop fix the real fault? Did it replace the part it billed for? Was the repair necessary at all? You usually can't know. The car runs, so you assume it worked.

This matters because a review is only as good as the reviewer's ability to judge. A restaurant diner can taste the food. A mechanic's customer can mostly judge the experience around the repair: how quickly the phone was answered, whether anyone was condescending, whether the final bill matched the quote, whether the car was clean when it came back.

Those things count. But they're a different thing from competence or honesty. A shop can be warm, fast, and wrong. A shop can also be curt, cluttered, and scrupulously right. Stars blur the two together.

Why the average is almost always high

There's a second distortion layered on top. Researchers who study online ratings, including Nan Hu, Paul Pavlou, and Jie Zhang in work on product reviews, have described a "J-shaped" distribution. Most ratings pile up at five stars, a smaller cluster sits at one star, and the middle is thin.

The reason is self-selection. People with a strongly good or strongly bad experience feel moved to post. People whose experience was merely fine, which is most people, stay quiet. The ratings you see aren't a sample of customers. They're a sample of the motivated.

For a mechanic, this creates a strange effect. A 4.8 average doesn't tell you much, because nearly every functioning business lands somewhere near it. The difference between 4.6 and 4.9 is often noise, shaped by how aggressively a shop asks happy customers to post. The number is a blunt instrument. The text underneath is where information lives.

What a useful review actually contains

The trick is to stop counting stars and start sorting sentences into two piles: the ones you can verify and the ones you can't.

Unverifiable praise: "Great service!" "Honest guys." "Highly recommend." "Fast and friendly."

Verifiable detail: "They quoted $480 for the front pads and rotors and the final bill was $480." "The check engine light came back; they re-diagnosed it free and fixed it." "They showed me the cracked boot on the axle and let me keep the old part." "They told me the other items on the inspection could wait six months."

The second kind describes behavior, not feeling. And one behavior stands out above the rest: a shop recommending less than it could have. Telling a customer "that can wait" is costly to fake and rare to invent. A review that mentions it is worth ten generic raves.

Other strong signals:

  • The quote matched the bill. This speaks directly to honesty on scope and price.
  • A problem came back and the shop handled it. Every shop has comebacks. How they respond tells you more than a clean record would.
  • The reviewer names a specific car and a specific repair. A review for a timing belt on a particular model is checkable against typical prices. "Fixed my car" isn't.

Reading the one-star reviews properly

Negative reviews deserve the same sorting, because they're distorted too. Some are pure venting. Some are about a bill the customer didn't understand. But a few contain the real warning signs.

Weigh these heavily:

  • Work done that was never authorized, or a final bill well above what was approved.
  • A repeated pattern across different reviewers, such as the same complaint about surprise fees or an inspection that always finds "urgent" items.
  • A shop that argues, rather than corrects, when someone has a problem.

Discount these:

  • Complaints about wait times, when the reviewer didn't book ahead.
  • Anger that a diagnostic fee exists at all.
  • A single furious reviewer with no specifics.

And read the owner's reply. A defensive reply that attacks the customer says something. A calm reply that explains the work order and offers to look again says something else.

Spotting reviews that were produced, not written

Fake and incentivized reviews exist in the auto repair space as they do everywhere. You can't always detect them, but a few patterns raise suspicion: a burst of near-identical five-star reviews within days, reviews that praise vaguely without mentioning a vehicle or repair, and reviewers with only one review in their history. None of these is proof. Treat them as reasons to rely on other evidence, not as verdicts.

Your next moves

  • Search the reviews for numbers. On any shop you're considering, scan for dollar amounts and phrases like "quoted" and "final bill." Read only those reviews first.
  • Hunt for the "it can wait" sentence. Look for reviewers who say the shop talked them out of work or deferred it. Mark any shop with several of these.
  • Read the owner replies on the three lowest-rated reviews. Note whether the shop explains or attacks.
  • Cross-check with a source that isn't stars. Look for a shop with technician certifications displayed, such as ASE, and check your state's consumer protection or licensing office for complaints. Neither is a guarantee, but they're not self-selected the way reviews are.
  • Test the shop with a small job before a big one. An oil change or tire rotation costs little. Ask for the written estimate, ask what they'd recommend, and see whether the answer matches the vehicle's manual. You're running your own low-stakes review.

The review you can write yourself

The lesson isn't that reviews are useless. It's that they answer a narrower question than the one you're asking. They tell you how a shop treats people. They rarely tell you whether the shop got it right. To close that gap, you need your own record: what your car was quoted, what you paid, what was done, and when.

That's the idea behind TrueQuote. It keeps your maintenance history in one place and helps you sanity-check a quote, such as whether $1,200 for brakes is in a fair range, so you're not relying only on a stranger's star rating or your own gut at the service counter. If you'd like a way to build that record, you can take a look at TrueQuote. Either way, read the next review for what it can prove, not how it makes you feel.