Facebook Ads for Solar Panels: How Crazygraph reduced lead costs

This case study demonstrates a successful lead generation campaign for a client in the highly competitive Australian solar panel industry. Our focus was on drastically reducing the cost per lead while maintaining a high quality of inquiries, showcasing the power of strategic Facebook advertising and continuous optimization. 

Our Goal

The primary objective was to significantly reduce the cost per lead for solar panel inquiries, ensuring a sustainable and efficient client acquisition process for our partner.

Strategy & Planning

Our approach to the Australian solar panel market involved a meticulous and data-driven strategy. Understanding the specific needs and pain points of potential solar customers in Australia was paramount. Our planning included:

Initial Data Collection & Analysis

Even without extensive prior data for this specific client, we leveraged industry benchmarks and initial broad targeting to quickly gather actionable insights.

Targeted Audience Segmentation

We developed detailed audience profiles, considering demographics, interests, and behaviors relevant to solar panel adoption in Australia. This allowed us to tailor our messaging precisely.

Compelling Ad Creative Development

Working closely with the client, we crafted visually appealing and benefit-driven ad creatives that resonated with the target audience, highlighting the advantages of solar energy.

A/B Testing of Ad Sets & Creatives

Continuous testing was fundamental. We systematically tested different ad copies, visuals, calls-to-action, and audience segments to identify the most effective combinations.

Rigorous Optimization

Daily monitoring of campaign performance, lead quality, and cost metrics allowed for real-time adjustments and budget reallocation to top-performing ad sets.

 


 

Results

The optimization journey, while challenging, yielded remarkable improvements in lead cost efficiency.

Initial Cost Reduction Phase

As demonstrated above, our initial strategic adjustments led to a significant improvement, bringing the cost per lead down from $175 to $151. This initial success provided a strong foundation for further optimization.

Advanced Optimization & Further Cost Reduction

Through relentless A/B testing and granular optimization, we achieved even more substantial reductions. We successfully took lead costs that were initially as high as $832 and $238 and brought them down to an impressive $160 and $201, respectively. This demonstrates a significant increase in efficiency and a strong return on ad spend for our client.

 


 

Conclusion

This case study emphatically proves that with a precise blend of strategic planning, targeted execution, and ongoing optimization, even highly competitive industries like solar panel installation can achieve outstanding results with Facebook Ads.

By systematically driving down the cost per lead from initial high points to efficient figures of $151, $160, and $201, we delivered high-quality leads and established a scalable framework for client acquisition in the Australian market. This success underscores the power of data-driven marketing to transform advertising spend into tangible business growth.

Ready to optimize your lead generation and scale your business? Book a meeting today, work with Crazygraph, and make your brand impossible to ignore!

Why our clients love
working with us!

Related Case Studies

Ready to take your
business to the next level?

Methodology

How we measure AI visibility

What every number in Beaconly means, how it’s calculated, and where its limits are.

Which AI we ask

  • Gemini, through Google’s official Gemini API. Right now Gemini answers from its own knowledge, without live web search, so answers don’t list sources.

  • Every question is asked as a fresh conversation, with no memory or personal history, set to the market (country) you chose.

What we ask

The questions a buyer would ask when looking for what you sell — like “best CRM for a small agency”. We write starter questions from your website; you can edit, delete or add your own. Questions don’t include your brand name, so we measure whether AI brings you up on its own.

Mentioned

Your brand counts as mentioned when its name, one of its other names you added, or its website address appears in the answer as a whole word. The same rule applies to every competitor, so the comparison is fair.

Position

The order in which tracked brands are first named in the answer: 1 means you were named before any tracked competitor. We show how often you’re named first, 2nd–3rd, later, or not at all.

Recommended

Mentioned and described positively. A separate AI step reads each answer and rates how it talks about each brand: positive, neutral or negative.

Visibility and small samples

Visibility is the share of answers that mention you. With fewer than 20 answers a percentage can mislead (1 answer changes it by 20 points), so we show the count instead — for example “3 of 5”.

Perception scores

We ask the AI to compare you with your competitors on the traits you choose, then score what the answer says about each brand from 1 to 10, using only the answer — not our own opinion. “What AI says about you” groups the reasons AI recommends you, its criticisms, and who it prefers instead, counted across your latest tracked answers.

Brand facts check

We ask the AI about your brand directly and compare each factual claim with the facts you gave us. Claims your facts don’t cover are marked “not covered”, never “wrong”.

AI traffic

The snippet on your site counts visits whose referrer is an AI assistant (ChatGPT, Perplexity, Gemini, Copilot, Claude and others), or that carry ChatGPT’s utm_source tag. Some apps hide the referrer, so the real number is usually higher. No cookies are used and no IP addresses are kept.

Page watch

We re-read the pages you watch once a day and look for your brand name, other names, or website address in the page text. Some sites block automated visits; we tell you when we can’t read a page.

Answers vary

AI assistants give different answers to the same question, and API answers can differ from what a signed-in person sees in the app (memory, location, model updates). That’s why we track over time and link every question so you can ask it yourself.

Spotted something wrong?

Email [email protected] with the page or brand, and we’ll look into it.