Managed cloud data infrastructure and services for SMEs and enterprises. Posts about data and analytics.
Joined November 2016
- Tweets6.8K
- Following729
- Followers3.6K
- Likes156K
Pinned Tweet
Having issues with the 1M event limit in GA4? Look into Parallel Tracking! #measure #ga4 buff.ly/3P4DXa8
GA4's native export goes to one place: BigQuery.
On Snowflake, Redshift or S3? The usual route is GA4 → BigQuery → your warehouse: two platforms to pay for, a 1M events/day cap, and a load that lands when it lands.
There's a shorter route.
reflectivedata.com/how-to-ex…
Warehouse bills rarely grow because of storage. They grow because every dashboard scans raw event tables each time it's opened.
Model once. Aggregate once. Point reports at the small tables.
One client cut marketing warehouse costs ~75% this way.
reflectivedata.com/case-stud…
3/ Shortcut 2: point it at raw warehouse exports. Complete data, but large, nested and full of duplicates, bots and test traffic. The model guesses the business logic on every query. Slow, expensive, inconsistent.
4/ What works: business logic in dbt, metrics in a semantic layer, and the AI reads from that. Same question, same number, every time. How we do it: reflectivedata.com/reflectiv…
dbt in production, a short checklist:
› every model in version control
› tests that fail before the dashboard does
› docs people can actually find
› schedules that don't live on someone's laptop
Missing two or more? That's what hosted dbt is for.
reflectivedata.com/platform/…
Pattern we see in most analytics audits: the tracking is fine, the definitions aren't.
"Conversion" in GA4, in the ad platform and in finance are three different numbers.
Nobody is wrong. Nobody agreed.
reflectivedata.com/services/…
4/ Cost: GA360 is an enterprise contract. Parallel tracking is priced by volume and is far more affordable for most teams that only need the data.
5/ Not sure which one you are? We'll tell you straight, even if the answer is GA360. reflectivedata.com/ga4-pain-…
AI answers about your data are only as good as the definitions under them.
If "revenue" means three things in three dashboards, your AI agent will confidently pick one.
Fix the semantic layer first. Then point the AI at it.
reflectivedata.com/platform/…
A dashboard nobody opens is a cost, not an asset.
Before building another one:
1. Who opens it, and how often?
2. Which decision does it change?
3. Does every metric on it have one agreed definition?
If #3 is no, start there.
reflectivedata.com/services/…
Things data engineers should not be doing at 2am:
› restarting a stuck connector
› backfilling a missed day by hand
› explaining to marketing why yesterday is empty
Managed pipelines exist so you can stop babysitting them.
reflectivedata.com/data-pipe…
5/ Option 2: GA360. Higher limits plus the rest of the 360 feature set, priced as an enterprise contract.
6/ Option 3: send a second copy of every GA4 hit straight to your own warehouse, unsampled and bot-filtered. That's GA4 Parallel Tracking. Full write-up: reflectivedata.com/how-to-av…
![How to Avoid Google Analytics 4 [GA4] BigQuery Export Quota Limit of 1 Million Hits - Reflective...](https://pbs.twimg.com/card_img/2115486904112369665/-IAeQifo?format=jpg&name=800x419)