Hey everyone! 👋 I’m pretty new to the whole ad ops/data side of things, but I’ve been trying to build up my analytics skills. I work with a lot of cross-platform data (mostly Meta and Google Ads right now) and was getting super frustrated trying to manually compare creative costs.
So, I built a Google Apps Script tool in Sheets to pull everything together! It’s probably super basic for the experts here, but I was really excited to get it working. Basically, it fetches ad performance data from both platforms (via their APIs) into one sheet, normalizes the cost metrics, and generates a simple report showing cost per acquisition by creative asset across both channels.
My main goal was to stop jumping between tabs and spreadsheets. Now I can see at a glance which images or videos are performing efficiently, regardless of where they’re running.
Here’s what my script does in a nutshell:
* Sets up API connections (using OAuth for both platforms).
* Pulls in campaign data for a specified date range.
* Matches creative IDs/names to aggregate costs and conversions.
* Outputs a pivot table style view in a new tab.
I’m sure there are way more sophisticated tools out there (like Looker or even dedicated bid management platforms), but this was a fun learning project for me! I’d love any feedback or suggestions for improvement.
My questions for you all:
1. For those who do cross-platform reporting, is this a common pain point? How do you usually solve it?
2. Are there any specific metrics or dimensions you always include in creative cost reports that I might have missed?
3. If I wanted to take this further, maybe even towards a simple dashboard, would you recommend connecting Sheets to something like Data Studio/Looker Studio, or should I look into learning dbt for better data modeling?
Really appreciate any tips or resources you can share!
>My main goal was to stop jumping between tabs and spreadsheets
That's the right starting point. Automation that reduces toil is valid SRE work, even if it's for ads data.
But you're missing critical layers. Where's your error handling for API failures? What's your retry logic? Have you set up any alerts for when the data flow stops? A script that breaks silently creates more work, not less.
You need to instrument this. Log execution times, track row counts per run, and set up a dead man's switch.
Five nines? Prove it.