Google Ads Case Study: AC Repair — 692 Leads at UAH 81 ($1.8) and a 24% Conversion Rate

The project is air conditioner repair and maintenance in apartments and houses. Why is it interesting? When the client came for an estimate, I did not believe in the numbers we ended up getting.

Contents

1. What we started with
2. Tracking: three key events
3. Results over 110 days
4. Which campaigns work
5. Performance Max on 20% of the budget
6. Seasonality
7. Why the forecast was lower
8. FAQ

Google Ads case study: AC repair and maintenance, CPL UAH 81, CR 24%, 692 leads

What we started with

The ad account already existed. But tracking was set up incorrectly: there was no call tracking, and the campaign was optimizing for clicks on the phone number.

A click on the number is not a call. A person could tap by accident, change their mind or just copy the number. Such clicks are mostly unrelated to real requests, and Google Ads was learning from exactly them.

So when I made the estimate, the maximum conversion rate I assumed was 10%.

Tracking: three key events

The project now tracks three key events:

• a real call — via dynamic call tracking, i.e. swapping the number on the website to see which click the call came from;
• Get Call — a request through the callback widget;
• form submission — unique requests only.

Duplicates are excluded as much as possible, and tracking is 100% in order. The landing page is as simple as it gets. The vast majority of requests come by phone — especially urgent ones, i.e. repairs.

Analytics · Google AdsThree Months Treating the Ads. The Front Desk Was SickA clinic marks qualified calls in Binotel, but never the missed ones. Why Google Ads sees only half the leads — and why the problem isn’t the ads.Oct 16, 2026Read

Results over 110 days

Over 110 days of work (March 14 — July 1, 2025):

• 692 leads for AC repair and maintenance;
• UAH 81 ($1.8) per lead;
• 24.09% — conversion rate;
• UAH 56.1K (~$1,250) — ad spend.

Instead of the forecast maximum of 10% — 24%. I would never have believed it was possible if I had not seen it with my own eyes and set it up myself.

Which campaigns work

For a long time I worked only with Search campaigns. They still have the highest conversion rate — almost 37%.

At the first stage I also launched a demand-generation campaign. The experiment failed: I launched it outside peak season.

Then I experimented with bidding strategies, balancing the cost per lead between manual and automated ones. With automated strategies the cost per lead is hard to control. I also tried Target CPA — it did not deliver.

Manual strategies, on the other hand, work great in niches where it is important to keep the cost per lead within a limited budget. This approach worked, and the Search campaigns still run on manual bidding.

Google Ads case studyB2B Google Ads: 195 qualified leads in a year on a 146,000 UAH ($3.2K) budgetIndustrial heating case study: after an audit and a switch from Performance Max to Search — 195 leads in a year on 146,000 UAH ($3.2K).Aug 31, 2026Read

Performance Max on 20% of the budget

Later I launched Performance Max on a small budget — about 20% of the total. The client asked for it. I was against it: in lead generation, the quality of Performance Max leads is always lower than from Search.

The hypothesis was confirmed. But because the cost per request in Performance Max was lower, the final cost per qualified lead came out about the same as on Search. So I kept those 20% on Performance Max — it is still running, and it is fine.

I wrote in detail about when Performance Max suits lead generation and when it does not in a separate article.

Seasonality

I recorded this case in early July 2025, when the peak season had not even started yet. The peak comes when the temperature outside goes above 30°C.

People in this niche hardly plan ahead. They reach out at the last minute — when the air conditioner stops working in the heat. So all the numbers above are from before the seasonal peak.

Why the forecast was lower

Honesty and integrity are among my core values. That is why I do not show clients my best cases when making an estimate. I show average market figures — what can realistically be achieved in the niche.

If it turns out better — great: the result exceeds expectations. That is better than inflated expectations and broken promises.

Google AdsGoogle Ads Media Planning: How to Build a Realistic Forecast Before LaunchStructure of a pre-launch Google Ads media plan: assumptions, scenario-based ROAS forecast, site readiness checklist.Sep 4, 2026Read

Call tracking like in this case study is available as a separate service: call tracking for Google Ads and GA4.