A rep knocks, launches into the opener about how much the neighbours are saving, and the homeowner points at the roof. There are eleven panels on it. The list said no solar.
The list did not say no solar. The list said no record of solar, and somewhere between those two sentences is the entire subject of this page.
Four ways a “no solar” filter is wrong
Every one of these produces a home that has panels and reads as if it does not.
- The system was never permitted. Installations happen without a permit — small systems, DIY installations, jurisdictions with light enforcement. There is no record to find because no record was created.
- The jurisdiction’s permits are not in your file. Not every authority publishes machine-readable permit data, and coverage varies county by county for reasons that have nothing to do with solar adoption. A county with no solar permits in a dataset may be a county with no solar, or a county that does not publish.
- Nobody looked at the roof. Aerial imagery coverage is not universal. Roughly 4.8 million single-family homes in a national universe of 85,072,153 have no current roof classification — the difference between that universe and the 80,265,597 roofs that have been classified. Hawaii sits at 22.8% classified and Alaska at 0.1%. Those homes are unknown, not negative.
- The imagery is older than the array. Aerial imagery refreshes on a rolling basis. A system installed a few months ago may not be on the current photograph yet, which makes every imagery-based count a slight undercount of today.
The null is the whole problem
Databases have three answers to “does this house have solar”: yes, no, and nothing. The third one is not a value, and the single most common defect in a solar prospecting list is a query that treats it as one.
| What the record says | What it means | Safe to knock with a first-system pitch? |
|---|---|---|
| Classified, panels found | There is an array on this roof | No — different conversation entirely |
| Classified, no panels found | Someone looked and found nothing | Yes |
| No classification | Nobody has looked | Unknown — this is the row that burns reps |
| No solar permit on file | No approval was recorded, in whatever jurisdictions your source covers | Weak evidence, on its own |
Written as a filter, the correct version is classification exists AND panels not found. The wrong version, which is what most tools produce by default, is panels not found — and that quietly folds the entire third row into your list. This is the same error, in reverse, that makes most published rooftop-solar estimates too low: counting unevaluated homes as negatives.
So the question to put to any vendor is not “do you have a solar flag”. It is can you distinguish a checked negative from an unchecked record, and can I filter on that distinction. If the answer is no, the flag is a suppression hint rather than a filter, and should be treated accordingly.
Permits are the wrong tool for this direction
Permits are genuinely excellent for one thing this page cares about: dating a system. If you want to know whether an array is nine years old or two, the permit is the record that tells you. That is why how to find homes with solar panels leans on them for the upgrade and battery-retrofit cuts.
Absence is the direction they fail in. A permit record proves an event happened; the lack of one proves nothing about the roof, only something about the paperwork. And even where a solar permit does exist, it does not guarantee an array was ever built — the population of homes with a solar permit and no visible array is large enough that we wrote a whole page about it in solar permits that never got installed.
Which cuts both ways. It means permit-based suppression leaves systems on your list, and permit-based targeting puts non-existent systems on it. Neither error is visible in the file.
Then the hard part: it is almost everyone
Suppose you get the filter right. You now hold a list of about 75.8 million classified American single-family roofs with no panels on them. The national adoption rate is 5.55%, which means the no-solar filter removed one home in eighteen.
It is a suppression rule, not a targeting rule, and it stops being even that in the places where it matters most. In San Diego County, 42.0% of classified roofs carry panels — there, excluding them changes a territory meaningfully. In Alabama, where 99.87% of roofs have no array, the filter does nothing at all. The state-by-state spread is in homes with solar by state, and it is the difference between a filter that reshapes a route and one that is decorative.
So the honest structure of a solar prospecting list is: no-solar as the floor, and something else as the reason. The candidates are the ordinary ones, and they work because they are about the household rather than the absence of a product.
- Roof age. A homeowner about to replace a roof is about to make a decision an array would ride on top of. How to find homes with old roofs covers how that is inferred.
- Ownership tenure and recent purchase. New owners have no incumbent supplier and a stack of open decisions.
- Adoption on the same street. Neighbours with panels are the most persuasive fact available at a door, and it is a fact you can check before knocking rather than assert.
- Owner-occupancy. An absentee owner is a different pitch and frequently not a pitch at all.
What to ask a vendor
- What proportion of my target counties have a current roof classification? Not nationally — in these counties.
- Can I filter on “evaluated and negative” separately from “no record”?
- How old is the imagery behind the determination?
- Does the panel determination come from imagery, from permits, or from both — and which one wins when they disagree?
- Is there a confidence grade on the individual record, given that small arrays are the least reliable end of the distribution?
A vendor who answers all five is selling a measured negative. A vendor who answers none is selling the absence of a row in a table, which is not the same product and should not carry the same price.
Where Scout Data fits
Disclosure: Scout Data is our product, and this is a thing it does. List Builder holds a present-tense panel determination on U.S. single-family roofs from current aerial imagery, keeps unclassified homes separate from verified negatives, and joins the result to permit history, roof age, ownership and tenure — so the second signal is a filter rather than a second purchase. Phones are matched by name to the owner on title and scrubbed against the federal do-not-call registry before delivery. Scout puts the same determination on a canvasser’s map.
We do not claim it is complete. Roughly 4.8 million single-family homes have no current classification, and we would rather tell you which ones than quietly count them as having no solar.
Frequently asked questions
How do I find homes that do not have solar panels?
By requiring positive evidence that the roof was looked at and no panels were found — not by taking every home with no solar record. Those are different populations. The second one silently includes every home your source never evaluated, and in some counties that is most of them.
Can I just use solar permit data to exclude homes with solar?
Not on its own. A permit filter misses systems installed without a permit, misses jurisdictions whose permit records are not published or not digitised, and cannot tell you whether an approved system was ever built or has since been removed. Permits are excellent for dating a system and poor at proving one is absent.
How many homes in the US do not have solar?
Of the 80,265,597 single-family roofs classified from aerial imagery, 4,451,792 carry panels — so 75,813,805 classified roofs do not. That subtraction deliberately excludes the roughly 4.8 million single-family homes in the universe that have not been classified, because a home nobody has looked at is not a home without solar.
Is a “no solar” filter the same as a solar lead?
No, and treating it as one is how territories get burned. Around 94% of American single-family roofs have no panels, so “no solar” barely narrows anything. It is a suppression rule that keeps you from pitching a first system to someone who already has one — the reasons to knock have to come from somewhere else.