Chapter 1 of 12 · min 0:00 to 2:19
Why working with ad-level data from Meta and TikTok is so hard
The problem with ad-level data in Meta and TikTok Ads: slow APIs, massive row volumes, and tools like Google Sheets or Looker Studio that choke under the load.
Watch on YouTube · Spanish audio
The problem we see in the market
There’s a very specific problem with ad-level data. Digital marketing teams have a hard time getting it in real time and with full history, whether on Meta or TikTok Ads. It gets especially bad with many accounts: agencies with many clients, or teams that work across several countries and usually run one ad account per country.
Not having that data in one place and in real time creates a lot of friction in decision making: which ad is performing best, which one is bringing conversions and which one isn’t, how the experiments between ads turned out. That friction ends in under-optimized campaigns, because there’s no visibility to decide with confidence.

Why data access is so painful
The platforms, first of all, have no intuitive way to visualize this data, especially with many ads and many accounts. On top of that, the Meta and TikTok APIs take forever to return data: they break, they saturate, and even programmatic access is complex.
Volume is the other enemy
Ad-level data has the highest granularity in digital marketing: campaigns contain ad groups, and ad groups contain ads. That’s a lot of rows, especially when you need the breakdown by day or by hour.
At that volume, the usual analysis tools like Google Sheets or Looker Studio saturate: they’re not built for Big Data, they’re built for lighter analysis. Between the painful access and the huge volume, working with this data becomes really hard.
Frequently asked questions
What is ad-level data?
It's the most granular data in digital marketing: campaigns contain ad groups, and ad groups contain ads. Looking at metrics per individual ad, day by day, multiplies the row volume compared to looking at campaigns alone.
Why does Google Sheets choke on ad-level data?
Google Sheets and Looker Studio are built for light analysis, not Big Data. With many ads, many accounts and daily or hourly breakdowns, the volume goes beyond what these tools can handle without slowing down.
Does this problem only affect agencies?
No. It happens in agencies with many clients, but also in marketing teams that operate across several countries and usually run one ad account per country.
All chapters in this series
- 0:00 1. The problem: ad-level data
- 2:19 2. The solution: warehouse, history and incremental deduplication
- 5:15 3. A look at the final result in BigQuery
- 8:27 4. Step one: connect Claude to Detrics via MCP
- 11:02 5. Step two: explore the sources and design the data schema with Claude
- 13:51 6. How the Detrics warehouse works
- 19:31 7. The table groups created in Detrics
- 22:04 8. Step three: create the destination in your own Google Cloud
- 25:27 9. Step four: transfers and the historical load
- 30:53 10. The data in BigQuery: tables, rows and creatives in your bucket
- 33:19 11. Analyzing the data with Claude and the BigQuery MCP
- 37:41 12. Final step: the unified Meta + TikTok view
Want to build this with your own data?
