How Accurate Is Project Sunroof?

Can remote assessments of solar capacity really compare to getting an expert on your roof?

When Google's Project Sunroof—an online tool for assessing solar potential on your rooftop—launched last year in North Carolina, I quickly plugged in my address to see if going solar made sense. And I was surprised—and more than a little skeptical—of the results.

Despite the fact that I live in a tree-covered neighborhood with plenty of shade on my roof, it was reporting over 1,000 square feet of usable roof space, 1,624 hours of usable sunlight per year, and net savings of $6,000 over the 20-year lifetime of the system.

Many commenters shared my skepticism. They argued this was a tool to sell leads to solar companies and suggested I get a solar company out to my house to compare notes. So that's exactly what I did.

Having finally contacted my friends (disclosure: also former clients!) at Southern Energy Management, and had a little chuckle about the silliness of the Google Sunroof numbers, I arranged a time for them to come out. And then I checked back in with Google Sunroof. Lo-and-behold, in the 14 or so months since my first "assessment", Google appears to have updated its algorithms and/or fine-tuned its data, because as the screenshot above shows, it's now suggesting just 1,073 hours of sunlight and 423 square feet available.

That's quite a difference. And it's also suggesting a net loss of $4,135 over 20 years, as opposed to the $6,000 savings it predicted last year, and saying that my roof may not be best suited for solar panels.

So what do the experts say?

Graham Alexander, a Residential Solar Design specialist at Southern Energy—who put solar water heating on my previous house, and talked me out of solar electric back then—came out to take a look. Having climbed up on my roof, checked out the lay of the land, and taken some detailed measurements with a SunEye 210 shade measurement tool (he also has a drone he can use to measure tree height), we sat down to discuss the options. What was interesting, says Graham, is that Google uses slightly different language than he might (what exactly does usable hours of sunlight mean?), but the results were not far off. Here were his recommendations:

System size: 5.8kW (Google Sunroof recommended 5.75 kW)
Annual production: 3,650kWh (Sunroof didn't give a detailed production estimate.)
Turnkey cost: $19,000 ($13,200 after tax credits)
20-year savings: $9,000
Net 20-year cost: $4,200

So Graham's estimate of lifetime costs was just $65 different to that of Google Sunroof—not bad at all, and a good sign that this will become an increasingly useful tool.

It was still helpful, of course, to get a professional opinion and significantly more nuanced detail. Graham was able to tell me, for example, that a similar roof with no shade would have produced 8,300kWh with the same system—more than twice the efficiency—and advocated strongly that my dollars are better spent elsewhere in supporting renewables. (Thank goodness for honest sales people!) I also learned, interestingly enough, that assuming I had the roof space, the annual production on my north roof would actually be similar to the south. (South has much higher shade but better orientation.)

Based on a sample size of one, at least, Project Sunroof appears to be an increasingly accurate tool for getting a basic estimate about whether or not solar is worth exploring on your property. If it says "yes" or "maybe", then getting a more detailed quote from a reputable installer will be a must before signing on the dotted line. But if it says "no", you may have good reason to trust it. As Graham told me, Google Sunroof is probably best seen as a complement to, not a replacement for, in-person solar assessment and system design:

"Project Sunroof can be a very useful tool for homeowners to begin their research in determining a general cost/benefit for a Solar Energy system for their home, but should be used as an estimate only. I would allow a 15-20% margin of error on the initial estimate. Solar professionals can also use it as a preliminary tool in lead qualification to separate high-quality sites from more shaded locations, but will still need a site visit to determine true system efficiency as a 15% error can dramatically change the cash flow over the estimated 30 year life of a system."

Luckily, community solar may finally be a thing in North Carolina soon. So I'll just wait to put my money where the sunshine is.