Full text, not stars
The score is built from what reviewers actually describe: the events, the numbers, the outcomes. Not the star they attached. Stars summarize a mood; the text holds the facts.
Best solar companies The Solar.com Installer Score
We read every review. Here's exactly how we score them.
The Solar.com Installer Score grades solar companies 0-10 from the full text of every published customer review we can collect. Not a star average, not a survey, and never for sale. This page explains everything the score considers, end to end.
Everypublished review read in full, not sampled
6categories scored separately, 0-10
20+reviews required before any score exists
$0what a company can pay to change it
Step 1 · Collect
Start with the complete review record
For each company we collect every published customer review from major public review platforms: full text, star rating, date, and reviewer history. All of it, not the first page. The review counts shown next to each score tell you exactly how much material that score is built on.Multiple platforms
Reviews are collected across public platforms, such as Google and Yelp, so a company can't look good by pointing you at its best channel. Each platform's star rating and count are shown alongside the score for transparency.
Clean sources only
No competitor marketplaces, no installer-curated testimonial pages, no purchased review feeds. If a source can be gamed by the company being graded, it isn't a source.
Step 2 · Read
Only evidence counts
Every review is read and tagged for what it actually says, category by category. A review that just says “great company, five stars” contains no evidence about pricing, timelines, or workmanship, so it contributes nothing to those category scores. A review that describes a missed rebate deadline, a leak repaired within a week, or a crew that failed inspection twice is evidence, and it lands in exactly the categories it speaks to.Sales Integrity
How clearly and honestly the company explains the system, contract, savings, and incentives before you sign.
Price Transparency
How well the final price, financing terms, and incentive guidance match what customers were led to expect.
Communication
How reliably the company keeps customers informed and answers questions across sales, installation, and service.
Timeline Reliability
How well the company manages the path from contract through permits, installation, inspection, and turn-on — without the customer having to chase it.
Workmanship
The quality, professionalism, neatness, and property care of the physical installation.
Long-Term Support
How well the system performs after turn-on, how the company handles monitoring, warranty claims, and service calls — and its staying power to honor them, based on years active and verified install history in utility records.
Why six categories: “How was the company” is really six different questions, and companies are rarely equally good at all of them. Scoring them separately is what lets you see that a company sells honestly but runs late, or communicates poorly but builds beautifully. Those are the exact trade-offs a single star average hides.
Step 3 · Weigh
Not every review deserves the same vote
Once a review qualifies as evidence, its influence on the score depends on how credible, how substantive, and how recent it is. We publish the factors, not the exact weights; the score stays harder to manipulate that way. Directionally:Reviewer track record
A reviewer with a long history of reviewing other businesses carries more weight than a zero-history account created the day it posted. Platform trust signals like Google Local Guide status or an established Yelp social profile add credibility. Accounts that look manufactured count for very little.
Depth and specificity
A detailed account of a project (prices, dates, names, what went wrong and how it was fixed) outweighs a one-liner. Depth can strengthen a review's vote, but length alone never dominates: a long review with no substance is still a one-liner.
Recency
Reviews fade with age on a gradual curve. A complaint from 2019 about a sales team that has since turned over matters less than the same complaint from last quarter. But persistent patterns keep showing up in fresh reviews, so real problems never age out of the score.
Severity
Concrete harm moves a category score far more than mild praise or a minor gripe: financial loss, property damage, a system down for months, a company that stopped answering. The score is built to protect the next customer, so the worst outcomes weigh heaviest.
The star record, discounted
The company's full star history still matters as a secondary signal, since it's the broadest sample of customer sentiment. But it's credibility-weighted and age-discounted too, and it never outvotes what the review texts actually say.
Thin data gets no score
Under 20published reviews, a company shows “Not yet scored” rather than a number. And a small review base can't post an inflated score: with limited evidence, scores are pulled toward the market average until the evidence earns them away from it.
Step 4 · One exception
Long-Term Support also asks: will they be around?
Five of the six categories are computed purely from reviews. Long-Term Support is the exception, because its promise of monitoring help, warranty claims, and service calls in year six depends on the company still existing in year six.A modest durability signal
Verified installation records from public utility interconnection filings add a staying-power signal to this one category: how long the company has been building, and whether it's still building now. The effect on the overall score is deliberately small. Solar history is full of big companies that vanished, so scale can nudge this category but never carry it.
The zombie rule
A company with no recent installation activity in the public record is capped in this category no matter how good its old reviews were. Five-star warranty stories from 2020 are worthless if nobody answers the phone in 2026, and California has already lived this story more than once.
No data, no penalty
Companies that build mostly outside the public record (municipal-utility territory or other states) keep a purely review-based Long-Term Support score. We don't fabricate a durability signal we can't verify.
The rules
Fairness isn't a feature. It's the whole product.
- One pipeline for everyone. Every company in a market is scored by the same process, the same analysis, and the same rules, in the same run. There are no special cases.
- Never for sale. No company can pay for a score, a rank, or placement, and solar.com partners get no favorable treatment. Some of the companies we work with carry middling scores on our own pages. That's the point.
- No hand edits, ever. Scores come out of the pipeline and go onto the page. Nobody at solar.com can adjust a number.
- Companies can respond. Installers can claim their profile and flag factual errors at partners@solar.com. Corrections fix source data and re-run the pipeline; they never touch a score directly.
- Dated and versioned. Every score shows when it was computed, and refreshes re-score the whole market together so no company is graded under different rules than its neighbors.
What the score is not
It is not a star average. It is not an install-volume ranking; that's a separate, clearly-labeled column built from public records. It is not a list of who advertises with anyone. And it is not a verdict: a below-average category is the question to ask that company before you sign, not necessarily a reason to walk away.
0-10score scale, one decimal, computed, never assigned
6category scores under every headline number
FAQ
Can a solar company pay to improve its Solar.com Score?
No. There is no way to pay for a score, a rank, or placement. Scores are computed by one pipeline that runs identically for every company, and no one at solar.com can hand-edit a score. Companies can flag factual errors for correction, but corrections go through the same pipeline, never around it.
Why does a company with a 4.7-star average have a middling score?
Star averages count every five-star "great company!" one-liner at full value. The Installer Score doesn't: reviews with no concrete detail carry no category evidence, and specific, credible, recent accounts carry the most weight. A company can hold a strong star average while its reviews tell a weaker story about pricing surprises or timeline problems. That gap is exactly what the score is built to surface.
Why is a company marked "Not yet scored"?
Two reasons: the company has fewer than 20 published reviews (too thin a record to grade fairly), or its full review record is still being processed. We'd rather show no score than a score built on a handful of reviews.
Where do the reviews come from?
Published customer reviews on major public platforms, such as Google and Yelp, collected in full: every review, not a sample. Review counts shown with each score tell you exactly how much material the score is built on. We never use competitor marketplaces or installer self-published compilations as sources.
How often are scores updated?
Scores are refreshed as new reviews are published and each refresh re-runs the entire market under the same version of the methodology, so every company is always graded by the same rules. Each score carries its computation date.
Do install volume or company size affect the score?
Almost entirely no. Five of the six categories are computed purely from what reviewers report. The one exception is Long-Term Support, where verified installation records add a modest durability signal, because warranty support requires the company to still exist. Size alone can nudge that one category; it can never carry a score, and a big install count never offsets bad reviews.
Get bids from installers competing through solar.com.
Run your address and see direct pricing in about two minutes. No sales call required.
Methodology: Counts are residential PV interconnections in investor-owned-utility territory (PG&E, SCE, SDG&E) per CPUC Rule 21 public records through May 2026. Postal and name variants are merged. Municipal utility territory such as LADWP, Glendale, Pasadena, Burbank, Anaheim, and Riverside Public Utilities is not included in those CPUC counts. Counts show recorded installations, not quality ratings. Reported costs are self-reported on owned-system interconnections. No installer can pay for placement.