Reliability stats for the Model 3 -- structured quarterly survey

[mkaresh]

Since 2005 I've been conducting a car reliability survey designed to provide up-to-date, objective stats. The survey is quarterly, for frequent updates. It asks that all repairs be reported, but excludes any problems that first appeared before someone joined, to minimize bias. We also don't count problems fixed entirely with software.

I don't have a bias for or against any car. My top priority is providing the best possible reliability information, whatever the stats might be.

We've been providing stats for the Model S and Model X for a while now, partly due to help from members of this forum. Though the Tesla's repair frequencies are still high, they've improved considerably from earlier.

I'd like to do the same for the Model 3, now that a significant number are being delivered. Those who are interested in helping can sign up here:

www.truedelta.com/join

 

[jaguar36]

Seems like you're going to get rather biased information, since folks with issues are far more likely to report them than folks with no issues.

 

[mkaresh]

That is what happens with the typical online survey. This survey is designed much differently.

As mentioned in my post, we do not count any problem that first appeared before someone signed up to participate in the survey. Instead, people respond whether or not they've had a repair each quarter going forward, starting with when they join.

We have quite a few models in the survey for which hardly any owners have reported a repair. To pick an extreme example, with the 2016 Honda CR-V we have zero repairs report for 62 participating cars. (The all-new 2017 CR-V has not been nearly as problem-free.) This would not happen with the usual survey process, which is a one-time thing and asks if there have been past problems.

Because of how our process is designed, people ideally sign up before they've even gotten a car, as this is the only way problems noted at delivery will be included in our stats. (Any car can be enrolled, not just Teslas.) But it's okay to start at any point.

 

[mkaresh]

Ten Model 3s signed up so far. A good start, given how few have been delivered so far, but more is always better, and I'd like to have 40+.

 

[cwerdna]

Seems like you're going to get rather biased information, since folks with issues are far more likely to report them than folks with no issues.

If one is going to compare between all automakers, if your claim is true, wouldn't it be pretty safe to assume that most/all automakers will have the same latter issue with the responses?

 

[mkaresh]

While some forms of bias can be expected to even out across all makes and models, if you expect this to happen all on its own you'd better have very large sample sizes. It's best to try to reduce the bias in the first place, which we do.

It's also possible to have a survey process that's so flawed that even a very large sample size won't help. This is far too common, and people need to look into how a survey was structured and conducted before giving weight to the results. But of course hardly anyone does, and we have the social media-fed mess we do

 

[mkaresh]

With the next round about the begin, eleven Model 3s are signed up. A good start, but more will be needed to provide stats.

To help make this happen:

www.truedelta.com/join

 

[mkaresh]

Now up to 16. Minimum sample size is 25. How soon can we get there, and beyond?

 

[mkaresh]

We conducted a special survey that included the Teslas in January, covering owner experiences through the end of 2018.

When data from the second half of 2018 are added, the 2017-2018 Model 3 improved from a calculated stat of 111 repair trips per 100 cars per year to 88. The latter is still high, over three times the average, but within the range we tend to see with upscale European cars.

More promising: Tesla Model 3s sold in the latter half of 2018 with dual motors and other new features scored much better, 41 repair trips per 100 cars. We have a smaller sample size for these, 18 cars, but the improvement is so dramatic that it’s almost certainly a real one and not sampling error. In other words, Tesla might be getting its stuff together, at least on the reliability front.

On the site, these updated stats are only available to members.

In general it is no longer free to join the site. But since we might continue to survey Teslas, I'm thinking of letting Tesla owners join for free. If you're interested in this, PM me here or contact me through the contact form on the site, www.truedelta.com.

 

[vickh]

do you have older Model S data, say 2015-6 builds? Reliability comparable to Toyotas/Lexus?

 

[Big-T]

I'd be happy to post data to help out the project, but I don't want to pay for a subscription. If there's a free way to submit survey questions let me know.

 

[ChrisH]

We conducted a special survey that included the Teslas in January, covering owner experiences through the end of 2018.

When data from the second half of 2018 are added, the 2017-2018 Model 3 improved from a calculated stat of 111 repair trips per 100 cars per year to 88. The latter is still high, over three times the average, but within the range we tend to see with upscale European cars.

More promising: Tesla Model 3s sold in the latter half of 2018 with dual motors and other new features scored much better, 41 repair trips per 100 cars. We have a smaller sample size for these, 18 cars, but the improvement is so dramatic that it’s almost certainly a real one and not sampling error. In other words, Tesla might be getting its stuff together, at least on the reliability front.

On the site, these updated stats are only available to members.

In general it is no longer free to join the site. But since we might continue to survey Teslas, I'm thinking of letting Tesla owners join for free. If you're interested in this, PM me here or contact me through the contact form on the site, www.truedelta.com.

I participated in your survey when I had my Lancer. I thought you stopped doing these surveys?

 

[Big-T]

I participated in your survey when I had my Lancer. I thought you stopped doing these surveys?

You had a Lancer too? Small world, I had a bright blue '08 - died last year, but it was a solid car. How about you?

 

[mkaresh]

do you have older Model S data, say 2015-6 builds? Reliability comparable to Toyotas/Lexus?

I posted our latest results for the Model S in that forum here. Definitely not up to Toyota / Lexus, but the 2016.5-2017 is close to the average car of that age. The earlier ones aren't close.

 

[mkaresh]

I participated in your survey when I had my Lancer. I thought you stopped doing these surveys?

We've stopped for nearly all cars, with the final survey for them last July. I'd start it up again if a suitable partner or acquirer came forward. It became too much for me to manage myself--especially some stuff beyond the survey itself.

I did conduct a January survey for Teslas and some other models, about one percent of the total cars signed up. A manageable amount of work. Might do it again. Trying to see if we can get more Tesla owners involved, especially for recent 3s.

 

[ChrisH]

You had a Lancer too? Small world, I had a bright blue '08 - died last year, but it was a solid car. How about you?

Silver 08 GTS. Was so preseantly surprised how well it treated me. Had to replace an oil pan at 8 years of ownership and that was the only non maintenance repair I ever had to do in the ten years I owned it. Then I paid over double what I paid for that for my LR RWD model 3!!!! So far so good. Hope to get 10 great years out of this one too!

How did yours die? Did you replace it with a model 3?

 

[Big-T]

Yeah I had zero non maintenance repair for 9+ years, then one day I lost blinkers, lights all but high speed wipers + some other electrical stuff all at once, but it was intermittent. Long story short after many hours of diagnostics it turned out to be the main computer...$1,500ish to order and replace. That was too much money to sink into a car we weren't planning to keep.

It was always the plan to trade in the Lancer when the SR Model 3 became available, unfortunately the SR wasn't available last year when the problem arose so we had to play a little car swap game.

• Trade in '08 Lancer - New Honda CR-V for wife.
• I'm got the wife's '09 Escape
• Blue SR+ reservation in (last minute switch from MSM), delivery in the next couple weeks.
• '09 Escape to be sold after

 

[ChrisH]

Yeah I had zero non maintenance repair for 9+ years, then one day I lost blinkers, lights all but high speed wipers + some other electrical stuff all at once, but it was intermittent. Long story short after many hours of diagnostics it turned out to be the main computer...$1,500ish to order and replace. That was too much money to sink into a car we weren't planning to keep.

It was always the plan to trade in the Lancer when the SR Model 3 became available, unfortunately the SR wasn't available last year when the problem arose so we had to play a little car swap game.

• Trade in '08 Lancer - New Honda CR-V for wife.
• I'm got the wife's '09 Escape
• Blue SR+ reservation in (last minute switch from MSM), delivery in the next couple weeks.
• '09 Escape to be sold after

Lol OMG. We traded my lancer in on my wife’s 2018 CRV EX-L AWD. I drove her Hyundai Elantra for about 8 months. Plan was to wait for SR but was worried it would never come out and wanted the full tax credit so I got a LR RWD in MSM. Ordered in early August and was told 2-4 months (just in time for tax credit). Got the car in 18 days instead.

 

[jat255]

We conducted a special survey that included the Teslas in January, covering owner experiences through the end of 2018.

When data from the second half of 2018 are added, the 2017-2018 Model 3 improved from a calculated stat of 111 repair trips per 100 cars per year to 88. The latter is still high, over three times the average, but within the range we tend to see with upscale European cars.

More promising: Tesla Model 3s sold in the latter half of 2018 with dual motors and other new features scored much better, 41 repair trips per 100 cars. We have a smaller sample size for these, 18 cars, but the improvement is so dramatic that it’s almost certainly a real one and not sampling error. In other words, Tesla might be getting its stuff together, at least on the reliability front.

On the site, these updated stats are only available to members.

In general it is no longer free to join the site. But since we might continue to survey Teslas, I'm thinking of letting Tesla owners join for free. If you're interested in this, PM me here or contact me through the contact form on the site, www.truedelta.com.

This looks interesting, but sorry, I'm not going to pay you to have the privilege of entering in my car's reliability data to build out your service. I get next to nothing from that interaction, so I don't see how it should cost me anything. If you wanted people to enter information, you'd be better off paying them in small Amazon gift cards or something.