The problem
Every recommendation started with a manual scan.
The Marketing Agency manages more than 200 ad accounts across Google Ads and Meta. The data was always there, but turning it into a clear view of performance, issues and next steps meant scanning each account by hand, comparing periods and writing up recommendations client by client. One thorough review took 30–45 minutes. Across a portfolio that size, it was the biggest bottleneck in the agency’s workflow.
The team already used Claude, but every analysis still began with gathering data from Meta Ads Manager, Google Ads, GA4, Looker Studio and spreadsheets.
- 200+ Google Ads and Meta accounts reviewed by hand
- 30–45 minutes for each thorough account review
- Data spread across ad platforms, GA4, Looker Studio and spreadsheets
- AI analysis still slowed down by manual data gathering
How the team works now
Client data now goes straight into the conversation.
GoMarble connects live Google Ads and Meta data directly to Claude, with no manual exports. The team asks questions against the data itself: what changed, why, and what to do next.
Account analysis in Claude
Multi-account performance scans and period-over-period comparisons run in one session, so the team can brainstorm, troubleshoot and validate ideas without switching between platforms and spreadsheets.
Automated portfolio health check
The agency built a recurring health check on top of GoMarble. It scans the whole portfolio every few days and flags the accounts that need attention.
Creative diagnosis on Paid Social
The team can quickly see which creatives and angles are working and which are fatiguing, then turn those findings into concrete next steps and new tests.
An e-commerce client moved budget from fatigue to growth.
For an anonymised e-commerce client, the team analysed roughly 90 days of combined Meta and Google Ads performance in a single session. About 30% of spend was going to a handful of underperforming ad sets and two fatiguing creatives. The agency recommended moving that budget to the top performers and launching three fresh creative angles. In the following month, ROAS improved by an estimated 15–20%. Before GoMarble, the same analysis would have taken several hours of manual work across platforms.
The result
The hours saved went back into better analysis.
The biggest outcome was operational efficiency: the team now analyses more client accounts in less time, with more data behind every recommendation. On accounts where the agency acts on insights surfaced with GoMarble, it typically sees results within one to two months, mostly from reallocating budget faster and cutting underperformers sooner.
GoMarble has cut the time we spend on manual account reviews by around 75%, saving us an estimated 180+ hours a month across our portfolio. That’s the equivalent of more than a full-time role.
Anders Pedersen
Co-founder, The Marketing Agency
