Case Studies: Google Ads and Tracking Results, With the Caveats

01

How Did a Bankruptcy Law Firm's Google Ads Cost per Conversion Fall Four Years Running?

A bankruptcy law firm's Google Ads cost per conversion fell every year for four years after CFS Marketing took over the account from another agency in 2022: from $61.44 in 2021 to $51.12 in 2025, 16.8% below the 2021 baseline.

Baseline

The firm came to us from a previous agency. In 2021, the last full year under that agency, the account converted 13.43% of clicks at a cost per conversion of $61.44. That year is the baseline for everything below.

What we did

• Took over the account in 2022 and managed it through 2025.
• Read the search terms report and built negative keyword lists to block searches from people who were never going to hire a bankruptcy lawyer. That work removed 40% of the low-intent traffic the account had been paying for.
• Reported in cost per conversion every year, and in 2024 checked the account's conversions against the firm's CRM.

Result

• Cost per conversion fell every year, closing 2025 16.8% below the 2021 baseline.
• 1,511 conversions from 2022 to 2025.
• 2024 had 475 conversions, verified against the firm's CRM.
• Conversion rate rose from 13.43% to 22.80% in the takeover year, then settled around 18%.

Honest caveat

Confidence: B (probable). The trend is consistent across four years and starts from a clean baseline year under another agency, but we can't isolate our work from everything else that moved in that window.

• The figures are Google Ads yearly figures. Only 2024's conversion count was checked against the firm's CRM. The other years are as Google Ads reported them.
• Conversion rate peaked in 2022 and came down. After the takeover year, the improvement is in cost per conversion, not in a rising conversion rate.
• Budget, auction prices and demand for bankruptcy help changed from year to year. We can't separate those effects from our management.
• This is one account in one market. It isn't a promise of the same outcome for your firm.

More detail on how we run legal accounts is on our law firm PPC and bankruptcy attorney marketing pages.

02

How Did a Trade School Go From 1,462 Phantom Conversions to 88 Verified Leads?

A trade school's Google Ads account reported 1,462 conversions, and when CFS Marketing cross-referenced them against the school's form database, zero matched a real lead. After we rebuilt the tracking, the account produced 88 verified leads from June to December 2025, and 27% of them became students.

Baseline

The account looked productive on paper: 1,462 conversions. None of them matched a real lead in the school's form database. Google's bidding had been learning from conversions that didn't exist, so every optimization built on that data was aimed at the wrong target.

What we did

• Audited the conversion tracking and compared every reported conversion with the school's form records.
• Rebuilt the tracking so a conversion meant a real inquiry.
• Ran Google Ads against the corrected data and checked each lead against the school's database.

Result

• 88 verified leads from June to December 2025, every one matched to the school's records.
• 27% of those leads became students.

Honest caveat

Confidence: B (probable) for the tracking finding; C (directional) for the lead and enrollment figures. The 1,462-to-zero mismatch is a direct database comparison. The lead figures cover seven months and 88 leads, which is a small sample.

• There is no clean "before" lead count to compare against, because the old tracking couldn't tell us how many real leads the account produced.
• Enrollment depends on the school's admissions process, not only on marketing.
• One school, one market, one season. Your numbers will differ.

The lesson applies in every industry: check what you're counting before you optimize it. Our GA4 and conversion tracking page covers the audit.

03

What Has Cesar Done In-House?

Before and alongside CFS Marketing, Cesar Soto held in-house marketing roles. These results come from employment, not client engagements, and we don't name the employers.

Hospital network, SEO Specialist (2019 to 2020)

• Optimized 2,000+ Google Business Profile locations.
• Set up 200 new listings.
• Managed 500 locations through Yext.

B2B facility services company, Director of Marketing (2026)

• Sole owner of SEO, AEO/GEO, content, paid search and web development.
• Eliminated 3,000+ broken redirects and 200+ duplicate pages in the first month, taking Ahrefs site health from 75% to 99%.
• Restructured the site around 11 service pages, updating 11,000+ internal links with zero post-crawl errors.
• Mobile PageSpeed 96+.

He also led marketing in-house at a commercial real estate brokerage (Marketing Director, 2020 to 2026) and worked at a personal injury law firm in 2018. We list those as experience only and publish no results from them. See how this experience applies on our healthcare and B2B SaaS and tech pages, or read more on the about page.

04

Why Are Client Names Withheld?

Client names are withheld because client work stays confidential by default, and we publish a name only with written permission. Until then, each client is described by type, such as "a bankruptcy law firm" or "a trade school."

Anonymizing means more than dropping the name. We also leave out ad spend, revenue, the city and any keyword list specific enough to identify the business. Employers are never named on this site.

05

How Are These Results Measured?

Every figure on this page comes from a client's own account or records, not from a sample, an estimate or another agency's work.

How we rate confidence:

• A (attributable): clean baseline, isolated change, result held over time with no competing explanation.
• B (probable): clean baseline and a clear result, with at least one factor we couldn't rule out.
• C (directional): the result is real, but the sample or window is too small to settle cause.

Neither case study is rated A. We'd rather tell you that than round up.

06

What Should You Take From Two Case Studies?

Two case studies show what our method has produced in two specific accounts. They don't show a typical result, because two results aren't a distribution. What they do show is how we work: we start from a dated baseline, check reported numbers against real records, and report what we can't prove alongside what we can.

If you want to see how the same approach would apply to your account, see our Google Ads consultant page or all industries.