I've seen some concern stemming from the poor reviews of Tesla Solar on SolarReviews.com. The average review was so low, it raised some red flags for me. As it stands right now, they rate Tesla at 1.5 stars out of 5 (and claim an average of $5.09 per watt when it's $2.01 as of writing).
I'm a Data Scientist, so I decided to do a bit of a meta-analysis of those 422 reviews. How have the reviews changed over time? They started more diverse, but became a lot more homogeneous after 2016. Dropping to unreasonably low levels as if a switch was tripped in 2018. Here are the average stars per review from 2014 to today:
In fact, if you do some basic text analysis of the written reviews, they started off using many different keywords, but in 2016 the reviews devolved into a lot of complaints purely about "customer service."
So what would suddenly affect the quality and score of these reviews? Around the same time, this page appeared on SolarReviews: Buying residential solar leads vs generating your own in 2020 According to the Internet Archive, that lead generation page appeared around mid-2016 (the gaps between 2018 and 2020 are due to them blocking robot traffic):
On that page, they clearly say that all three Solar related websites they own are "lead generation websites":
And they ask solar installers to pay for leads and accreditation on their website. So starting in 2016, Tesla reviews didn't exist on their website for the purpose of informing their visitors. They were hosting those reviews with the express purpose of directing web traffic to paying competitors.
It's hard to prove when reviews are faked, so all we can do is show the ratings over time and point out the structural shifts. If you're curious at taking a look at their reviews in aggregate yourself, I've attached a CSV of their reviews to the bottom of this post.
I'm a Data Scientist, so I decided to do a bit of a meta-analysis of those 422 reviews. How have the reviews changed over time? They started more diverse, but became a lot more homogeneous after 2016. Dropping to unreasonably low levels as if a switch was tripped in 2018. Here are the average stars per review from 2014 to today:
In fact, if you do some basic text analysis of the written reviews, they started off using many different keywords, but in 2016 the reviews devolved into a lot of complaints purely about "customer service."
So what would suddenly affect the quality and score of these reviews? Around the same time, this page appeared on SolarReviews: Buying residential solar leads vs generating your own in 2020 According to the Internet Archive, that lead generation page appeared around mid-2016 (the gaps between 2018 and 2020 are due to them blocking robot traffic):
On that page, they clearly say that all three Solar related websites they own are "lead generation websites":
And they ask solar installers to pay for leads and accreditation on their website. So starting in 2016, Tesla reviews didn't exist on their website for the purpose of informing their visitors. They were hosting those reviews with the express purpose of directing web traffic to paying competitors.
It's hard to prove when reviews are faked, so all we can do is show the ratings over time and point out the structural shifts. If you're curious at taking a look at their reviews in aggregate yourself, I've attached a CSV of their reviews to the bottom of this post.
Attachments
Last edited: