Solar leads are usually cheap but unconverting for one reason: Meta is doing exactly what it was asked to do, which is find people who will submit a form, not people who will sign a contract. When the form is easy, the offer is vague and the only signal the campaign receives is "lead submitted", the algorithm fills your pipeline with the cheapest form-fillers it can find. The fix is to change what counts as a lead, both on the form and in the signal you send back to Meta.
This article walks through how to locate the leak, how to calculate the number that matters (cost per installation, not cost per lead), and what we change first in solar accounts. We use our work with a solar and backup-power installer in Ukraine as a reference point.
Cheap leads are a signal problem first
A lead-form campaign optimizing for leads learns from one event: a completed form. It has no idea whether that person owns a roof, pays a large electricity bill or picks up the phone when your sales team calls. Every cheap lead that never converts still counts as a success in the campaign's eyes, and the algorithm goes looking for more people like them.
That is why "make the leads cheaper" and "make the leads better" often pull in opposite directions. A shorter form, a softer offer and broader creative will almost always lower cost per lead. They will also widen the gap between what Meta reports and what your installers see.
The useful question is not "why are my leads cheap?" but "where in the funnel do they stop?" The answer tells you which fix to apply.
Step 1: find where the leads stop
Pull the last 30 to 60 days of leads and tag every one with the furthest stage it reached. A simple ladder for solar looks like this:
- Lead submitted (form completed)
- Contacted (someone from your team spoke to them)
- Qualified (owns the property, roof or site is suitable, budget and timeline plausible)
- Site survey or technical visit booked
- Quote sent
- Contract signed
- Installation completed
Then look at the drop between each pair of stages. The pattern points to the cause:
- Large drop from submitted to contacted. People are not answering. Check speed to first call, whether the phone field is validated, and whether the creative promised something your call does not deliver.
- Large drop from contacted to qualified. The wrong people are filling in the form: renters, people in apartments without roof access, people curious about prices with no intention to buy. This is a form and offer problem.
- Large drop from qualified to survey or quote. Usually a sales process or timing issue: slow follow-up, unclear pricing, or a season when installs are hard to schedule.
- Large drop from quote to contract. Price, financing or trust. Ads cannot fix this alone, but they can set expectations earlier.
If you cannot build this ladder because your CRM does not record stages consistently, that is the first problem to solve. Without stages, you are optimizing blind and any agency, including us, is guessing.
Step 2: tighten the form without killing volume
Meta lead forms let you trade some volume for intent. Two settings matter most.
The first is the form type. The Higher intent option adds a review step where the person confirms their details before submitting. It removes a share of accidental and low-effort submissions. Expect cost per lead to rise; judge it on cost per qualified lead instead.
The second is the questions themselves. For this installer we built a multi-question qualifying form so casual enquiries filter themselves out before reaching the sales team. Useful solar questions include property type, whether the person owns the property, approximate monthly electricity bill, and what they want: grid-tied panels, backup power or both.
One caution from that account: a strict pre-filter can reject good leads. At one point a bill-size threshold on the form screened out a lead that the client later confirmed was real. The lesson we took is that the form is a pre-filter, not the judge. The CRM status, set by a person who actually spoke to the lead, is the source of truth for "qualified". Use form answers to route and prioritize, and be careful about hard-rejecting on any single answer.
Step 3: show the price earlier
Many solar ads avoid prices because the final quote depends on the site. The result is a lead pool full of people who will drop out the moment they hear a number.
With the installer we rebuilt the account around price-led creatives, with prices shown on the ads themselves. A price on the creative works as a filter before the click: someone who is shocked by the number does not submit a form, and someone who submits already accepts the order of magnitude. It does not need to be a final quote. A "from" price for a defined system size is enough to set expectations.
We test these as variants (static images in 4:5 and 9:16) and keep the versions that produce qualified leads, not just leads. Our creative testing systems article covers how we structure that rotation.
Step 4: send Meta the outcome, not the form
This is the change that moves the most in our experience. If the campaign only ever hears "lead submitted", it will keep optimizing for submissions. If it hears "this lead became qualified", it can start learning which people resemble your real customers.
Meta supports this through the Conversions API: your CRM sends lead-stage events back, matched to the original lead. Once enough of those events flow, a campaign can be set up to optimize toward leads more likely to reach the stage you choose, rather than toward raw submissions. For the installer we feed CRM-qualified leads back to Meta as the optimization signal.
Two rules keep that signal clean:
- Only send real qualified leads. Test submissions, duplicates and junk should never be sent as qualified events. One wrong event teaches the algorithm the wrong lesson.
- Use the stage your team sets consistently. If "qualified" means different things to different salespeople, the signal is noise. Write the definition down.
We explain the stage ladder and the mechanics in more depth in Meta Ads leads vs qualified leads for home service companies.
Step 5: sell what can be installed this season
Solar demand and installation capacity are both seasonal. In winter, rooftop panel installs stall in many climates. Leads generated for panels in that period sit in the pipeline, go cold and look like a conversion problem when they are really a scheduling problem.
For the installer's winter period we moved the offer to backup power (inverter plus battery), which can be installed and is relevant when outages and short days are on people's minds. Matching the offer to what the team can deliver now keeps the gap between "lead" and "installation" short.
A worked calculation: cost per installation
The following is illustrative arithmetic, not a client result.
Say you pay $3 per lead. You buy 100 leads for $300. Of those, 5 book a site survey, and roughly 1 in 3 surveys becomes an installation. That is about 1.7 installations, or roughly $180 per installation.
Now say a stricter form, a price on the creative and a qualified-lead signal push cost per lead to $8. You buy 100 leads for $800. This time 25 book a survey, and the same 1 in 3 close. That is about 8.3 installations, or roughly $96 per installation.
The second scenario costs more than twice as much per lead and about half as much per installation. It also wastes far less of your sales team's time: 100 calls produce 25 surveys instead of 5.
Run this with your own numbers. You need four inputs: spend, leads, surveys booked and installations. If your cost per installation is acceptable, cheap leads are not a problem. If it is not, the stage where the drop is largest tells you where to work.
Checklist: cheap solar leads that do not convert
- Every lead in the CRM has a stage, updated by the person who spoke to them.
- You know your survey rate and close rate from the last 30 to 60 days.
- First contact happens quickly after submission, ideally within the same working hour.
- The lead form uses qualifying questions: ownership, property type, bill range, interest.
- No single form answer hard-rejects a lead without a human check.
- Ads show a price or price range for a defined system.
- CRM-qualified leads are sent back to Meta through the Conversions API.
- Test, duplicate and junk leads are excluded from that signal.
- The current offer matches what your team can install this season.
- Reporting shows cost per qualified lead and cost per installation, not only cost per lead.
How we work on solar accounts
We start with the CRM, not the ad account. Before changing creatives we want to see the stage ladder and where it breaks. Then we work in this order: clean the signal, tighten the form, put prices on the creative, align the offer with the season, and only then scale budget.
For reference, in the installer's first 16 days the rebuilt account produced 144 lead-form leads on a spend of 19,147 UAH, an average of about 133 UAH (about $3) per lead. We report that figure because it is the one Meta measures; the work after launch is judged on how many of those leads the client's team qualifies. No setup guarantees a close rate, because that depends on your sales process and pricing as much as on the ads.
If this matches what you are seeing, see how we run Meta Ads for lead generation, or apply here. We reply in writing with what we would change first in your account. Pricing is fixed: the 14-day Growth Sprint is $1,199–1,299 and Managed Meta Ads start at $1,990 a month, confirmed in writing before you pay; the ad budget is paid separately, directly to Meta.
Frequently Asked Questions
What is a good cost per lead for solar on Meta?
There is no universal figure, and comparing cost per lead across markets or offers is misleading. A better target is the cost per installation you can afford given your margin. Work backwards from that using your survey rate and close rate, and the acceptable cost per lead follows.
Should I use Meta lead forms or a landing page for solar?
Both can work. Lead forms remove friction and usually produce more volume at a lower cost per lead, which is also why they need qualifying questions and a clean feedback signal. A landing page adds friction but lets you explain pricing and system options in more depth. Our [landing page anatomy](../landing-page-anatomy/) article covers what that page needs.
Does the Higher intent form type reduce lead volume?
Usually yes, because it adds a review step before submission. The point is to remove low-effort submissions. Judge it on cost per qualified lead and cost per survey rather than on raw lead volume.
How many qualified leads does Meta need to optimize for them?
Meta needs a steady volume of stage events to learn from, and the exact requirement depends on the setup. If qualified events are rare, a common approach is to optimize for an earlier stage that happens more often, such as "contacted" or "qualified", rather than for signed contracts.
Why do my installers say the leads are bad when Meta shows good results?
Because Meta measures form submissions and your installers measure jobs. Both are telling the truth about different things. Once qualified-lead events flow back into Meta and your reports show cost per installation, the two views start to line up.