You can't use the complaint totals to directly compare different vehicles, right?

Right, you can't. Not unless you really know what you are doing. Chances are the various models you're comparing had very different annual sales. More vehicles on the road usually means more complaints, & not necessarily a more problematic vehicle.

The exception is usually you can compare one model year-to-year using straight complaint totals, because annual sales figures typically don't change very significantly. That's not always true, but the really bad model years tend to stand out regardless of fluctuations in sales.

However comparing total complaints across different makes/models will probably get you nowhere, except maybe an "F" for poor logic*. But hey, play with fire if you want to. It's best to use the "per vehicle" method below. * Unless you know the sales figures for each model, then you've got a flying chance for a "B+". See "Why don't you factor in vehicle sales/production data?" below.

Then what good is all your complaint data? Why the heck are the totals even on the graphs?

The totals are accurate & useful when you look at what the graphs are showing you, per vehicle. What stands out as defective, compared to the other categories for that same vehicle? That's easy to figure out. Then go forth & compare that relationship to other vehicles. When you evaluate one vehicle at a time, vehicle sales figures don't matter.

Here's an example —

2007 Honda Accord 2007 Hyundai Elantra 2007 Toyota Camry
The Accord has some categories with more complaints than others, but nothing hugely defective. You'd probably want to click on the bars near the top to find out how serious those brakes, body/paint etc problems are. Also the NHTSA data shows engine complaints as one of the problem areas, but again, nothing really stands out as a massive problem. The Elantra is squeaky clean with no real defect trends showing up (in our data, anyway). Does it matter how many Elantra's were sold relative to the Camry or Accord? Nope. Nothing stands out as being a problem. The 2007 Elantra is a pretty safe bet for a good solid trouble-free vehicle. The Camry clearly has major engine problems, with 12 times the engine complaints than the average. The Camry also shows some less-common issues with brakes, transmissions & possibly interior accessories. Does it matter how many Camry's were sold? Nope. Regardless of sales figures, engine complaints are clearly off the chart & that's what matters.

Why don't you factor in vehicle sales/production data?

We'd love to, but apparently no one has this data. We have looked for a source for this, for years. We've asked data companies & industry professionals. So far, no one has it available. If you know of a good source for make/model/year-specific sales or production data for the last 15-20 years, please drop us a line. One site that's close is GoodCarBadCar.net.