Clinic: ads audit and cleanup
A private clinic felt its ads had stopped working. I audited the account against GA4, call tracking and the CRM, found campaigns that were spending without bringing patients, and switched them off. I then focused the budget on the search campaigns that actually drove calls.
Next month, compared with the month before:
calls from ads: +137%
patients who came to an appointment: +212%
ad spend: −16%
cost per patient: −73%
1
38
I managed Google Ads for W.A.H., a veterinary practice with a reasonable advertising budget but an early challenge: reaching the audience that would engage with the practice and generate more calls and form leads.
Rather than rely on additional spend, I worked on keyword coverage, performance-based pruning, and audience and location targeting. I reviewed search trends to identify promising keywords that were missing from the campaign, paused underperformers against monthly ad-group benchmarks, and refined the search radius.
For February 19, 2025–August 17, 2026, Google Ads recorded 253 conversions from CA$9,354.96 in advertising spend, equivalent to CA$36.98 per recorded conversion. The clearest pattern in my analysis was the contribution from keywords that explicitly named the local area: they accounted for 52.4% of conversion credits while using 22.3% of keyword spend.
0
5
A live earnings dashboard built on Airtable. My invoicing ledger could not tell me whether the year was on pace, so I built a React dashboard on top of the Airtable base that shows pace to target, a year-end forecast, seasonality and FX impact, with no manual report building. Full case study: https://klaracondac.com/work/airtable-earnings-dashboard
1
35
MRR Movement — New, Expansion, Contraction & Churn tracked monthly for a B2B SaaS case study.
0
39
CRO audit with practical ideas. The same method I use to run conversion analysis in a large, regulated organisation, applied to your site.
How it works:
- Funnel data first: conversion and drop-off for each step of your key journeys in GA4, by device and channel
- Screen-by-screen review of each journey on desktop and mobile, with annotated screenshots
- Every finding backed by evidence (data plus screenshot), never opinion alone
- Findings split into quick wins you can ship now and hypotheses worth an A/B test
- Ideas ranked by impact and effort, so you know what to do first
- A test roadmap: hypothesis, change, primary metric and guardrail for each test
- You get a clear report your team can act on, with the first 3 tests ready to run.
Education center: CPA −49% while scaling
An education center (exam and university-admission prep) needed more students without paying more per lead. As part of the agency team I launched Google Ads and Yandex Direct; together with paid social, hot leads from paid ads grew by 500%. Then I rebuilt the Google Ads account around lead quality from the CRM: new structure, tighter match types and negatives, structured tests of audiences, ads and bids.
Over 3 months, with the budget up 28%:
CPA: −49%
conversions: ×2.5
CTR: ×2
1
58
I managed Google Ads for T.C.A.T., an accounting and tax practice working with a tight advertising budget. My challenge was to get the client comfortable with the onboarding process, then make approximately $400 $500 per month work in a competitive tax-season market.
I moved from Maximize Clicks to manual bidding and built an hourly optimization script using roughly two months of engagement data. My goal was to prioritize the periods when search users were more likely to engage, rather than treat every hour as equally valuable.
The account’s all-time Google Ads results recorded 77 conversions from $6,345.76 in advertising spend, at $82.41 per recorded conversion. Conversions were primarily phone calls and form leads. After tax season, the client increased the monthly budget from $400 to $800, a vote of confidence in the account I managed, rather than a percentage increase in advertising performance.
can AI search find and quote this online shop? Sample audit of a pre-launch art shop in Lisbon, using public data only. I checked AI crawler access, llms.txt, structured data and answer-ready content, then ran 10 real buyer queries to see who gets cited instead. Result: 11 findings ranked by severity with fixes, 0 of 10 queries citing the shop yet, one open niche nobody answers, and a 30-day plan.