Selected works
Improving ad performance across seven Google Ads campaigns using AI
I built an AI-assisted auditing workflow for more than seven simultaneous Google Ads campaigns, helping Gravity Wiz identify ad opportunities (and problems) earlier. The system combined ad-platform data with on-site behavior so weekly decisions reflected more than Google Ads first party data — it reflected actual campaign performance.
Context
At our peak, Gravity Wiz had 7+ Google Ads campaigns running simultaneously. I’m managing all of them, and have a list of other things to do. I built an AI-auditing system to keep on top of opportunities and problems that saved my time while improving campaign performance.
My role
I designed the audit process, connected the required data, and used the output to prioritize campaign changes.
Actions taken
- Built an AI campaign-auditing assistant.
- Connected the assistant to Google Ads and PostHog via MCP.
- Made campaign data such as click-through rate, ad performance, active tests, and Quality Score available to the assistant.
- Made on-site behaviour such as bounces, user paths, and conversions available alongside the ad data.
- Built weekly automated audits at two levels.
- Produced detailed reviews for individual campaigns: breaking down performance, runaway spend, high performing assets, tests, and how these are impacting actions on-site.
- Produced an account-wide summary across all active campaigns.
- Created a severity system that ranked findings from critical errors to lower-priority opportunities.
- Used the audit output to direct manual investigation and campaign changes.
Results
- Reduced runaway advertising spend by 40% across flagship campaign (with dozens of ad groups).
- Increased ROAS by 110% in flagship campaign.
- Created a clearer account-wide view of how campaign changes affected performance.
- Made critical errors easier to identify and address first.
Next selected work