Agent speed shortens your reading window
The faster changes ship, the less clean time sits between them. Agents compress exactly the quiet gaps that measurement depends on, unless you reserve them deliberately.
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 2 min read
Run the numbers for your store: the free attribution window calculator.
Measurement needs quiet periods, and agents are very good at removing them. The resource being consumed is uninterrupted time.
The mechanism
A causal read needs a window in which the thing you are measuring is the main thing that changed.
An agent shipping continuously means there is no quiet window. Every period contains several changes, so every read is confounded, no matter how much data accumulates.
More data does not fix it
This is the counterintuitive part. Confounding is not a sample size problem. Ten times the traffic through a window where four things changed still cannot tell you which of the four did it.
You can buy precision with volume. You cannot buy identification with volume.
What has to be reserved
| Reserve | Typical size |
|---|---|
| An untouched control set | Chosen before results exist |
| A do-not-ship period around known events | Sales, launches, peak season |
The third is the one most often skipped. Shipping a catalogue change into the week before a peak trading period guarantees you will never know what the change did, because the season will dominate everything.
The trade, stated plainly
Reserving quiet time costs throughput. It buys the ability to say which change worked, which is what makes the next change better than a guess.
A team shipping continuously with no reads is not moving faster. It is moving at the same speed with no steering.
The read
Where the window was too busy, the honest output is the label rather than a number.
Related answers
Frequently asked questions
Why does shipping faster make measurement harder?
A causal read needs a window where the thing being measured is the main thing that changed. Continuous shipping means every window contains several changes, so every read is confounded.Can more traffic compensate for a busy window?
No. Volume buys precision, not identification. Ten times the traffic through a window where four things changed still cannot say which of the four caused the result.What should I avoid shipping into?
Known events such as sales, launches and peak trading periods. A change shipped into peak season can never be separated from the season, so the information is lost regardless of how it performs.
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
- AttributionAttribution identifies user actions that contribute to a desired outcome and assigns value to each. It reveals which marketing touchpoints drive conversions.
- ConfoundingConfounding is a distortion of the estimated treatment effect when a third variable, a confounder, associates with both the treatment and the outcome. Causal inference methods control for confounding to isolate the true treatment effect.
- IncrementalityIncrementality is the part of an outcome, such as orders or revenue, that a marketing activity caused: the extra results compared with what would have happened without it. It is a causal effect, not the number of conversions credited to a channel.
- ShippingShipping transports goods from a source to a destination. In e-commerce, shipping costs and delivery times directly influence customer purchase decisions.