Move from manual uploads to continuous reads
Four stages from an export somebody remembers to a scheduled read that reproduces on demand, and the single test that catches most pipeline bugs.
Three platforms, three answers, one budget to defend. Your GA4 export already holds the answer, and the first finding is free and never leaves your browser. Get your first finding free
By Joris van Huët, Founder & CEOUpdated 3 min read
Four stages take you from an export somebody remembers to a scheduled read that reproduces on demand. The reproduction test at stage three is what separates a working pipeline from one that looks like it works.
The four stages
| Stage | What you build | The test |
|---|---|---|
| 1 | Frozen definitions, written down and dated | Two people build the same channel list from them |
| 2 | One manual read you would defend | You can state its interval and coverage unprompted |
| 3 | Automated ingestion on a schedule | Re-running last month returns last month's numbers |
| 4 | Delivery into where decisions are recorded | Somebody replies to it |
Stage 1
Write the lookback window, the channel grouping, the timezone and what counts as an order. Date the document. This is the least interesting stage and the one that determines whether stage three is meaningful, because a pipeline encodes whatever definition it was given.
Stage 2
Run one manual upload and sit with the output. The €99 one-time read on a Google Analytics export exists for this, refundable within 30 days, no questions asked.
Stage 3, and the test that matters
The usual culprits are a retention boundary that rolled forward, a timezone mismatch, or a platform that changed a default. The retention case is in the two-month retention trap.
Do not skip this. A pipeline that silently returns different history each run will corrupt every comparison you make from it, and the corruption is invisible.
Stage 4
Deliver into wherever budget decisions are recorded rather than into a tool. There is no native Slack app or Notion connector; the routes are compared in putting attribution results into Slack or Notion.
What to alert on once it runs
The job, not the numbers. Ingestion failed, delivery late, coverage below your floor. Never on the estimate moving, which it does by design.
The order is not optional
Teams that start at stage three end up with a reliable pipeline producing numbers nobody agreed the definition of. That is more expensive to unwind than to do properly, because by then reports exist and people have opinions about them.
Related answers
Frequently asked questions
How do I test an automated attribution pipeline?
Ask it for a past month you already have numbers for, and check they match. If they do not, something in the chain is reading a moving target, usually a retention boundary, a timezone mismatch, or a platform default that changed.What should I automate first?
Nothing, until definitions are frozen and one manual read has been produced and understood. A pipeline encodes whatever definition it was given, so automating before agreeing the definition just makes an unresolved argument reproducible.What replaces manual uploads?
Pro costs €299 a month, excluding VAT, and adds automated GA4 ingestion, continuous attribution updates, lifetime look-back, AI chat over your own results and budget alerts. The €99 read is a manual CSV upload by design.
Go deeper: Causal attribution, explained.
Sixty-second versions of these ideas: Causality Engine on YouTube Shorts.
Have an audience of ecommerce operators? Earn 20% recurring commission by recommending the €99 read.
Keep reading
Terms in this article
- AnalyticsAnalytics is the systematic computational analysis of data. It reveals customer behavior and measures campaign performance.
- AttributionAttribution identifies user actions that contribute to a desired outcome and assigns value to each. It reveals which marketing touchpoints drive conversions.
- Google AnalyticsGoogle Analytics is a web analytics service that tracks and reports website traffic.