Almost certainly.
This concerns those fortunate enough to be working with GA4 exports in BigQuery. The thing is, there have been recent changes in the billing settings for your datasets. Now you can choose between logical and physical storage models.
There's a significant cost difference between them. The physical storage model is exactly twice as expensive as the logical one. However, it provides a compression level that can vary depending on the dataset.
By default, the schema with so-called logical bytes is used, which means uncompressed data. All would be fine, but GA4 exports consist of a multitude of similar and identical rows. Among other things, this implies that the compression ratio when using the physical storage model can be very high.
So, the math is straightforward. The physical storage model is twice as expensive as the logical one. Therefore, if the compression ratio is higher than 2, switching to physical bytes becomes cost-effective for you.
I've come across several cases (especially from Luka Cempre - I used his screenshot) where using physical bytes for GA4 exports led to a significant cost reduction. Yes, indeed.
To find out the sizes of your dataset in logical and physical bytes, go to the details of your dataset and check the Storage Info tab.
To change storage settings:
ALTER SCHEMA <dataset-id> SET OPTIONS( storage_billing_model = 'physical');
By referring to INFORMATION_SCHEMA.TABLE_STORAGE_BY_PROJECT, you can estimate the approximate cost difference for a specific dataset. If you need a query, ask me in DM.
If you have an interesting use case for using physical bytes in BigQuery, be sure to let me know in the comments.
If you found this post helpful, please like and share it with your friends – it's crucial for me.
