Scaling, proof and the cloned page: 12 answers
Twelve answers grouped into the three problems the last week of ecommerce discussion keeps circling: what scaling does to the read, what makes a number survive finance, and how to measure a page an AI rebuilt in minutes.
By Joris van Huët, Founder & CEOUpdated 2 min read
Three problems, twelve answers, and one rule that decides all of them: a number without a comparison is not a measurement.
The last week of ecommerce discussion keeps circling three things. Each is a measurement problem wearing a different costume.
What scaling does to the read
A budget ladder is a sensible way to grow and a poor way to learn. Reported return can hold steady across every step while the marginal return falls the whole way.
What makes a number survive finance
A finding is defensible when the answers to the obvious objections are already in the document. If it needs you in the room, it is a presentation.
- How to defend an attribution finding
- Judging an attribution vendor on transparency
- The real cost of causal attribution software
- LTV, creative volume, and what breaks measurement
How to measure a page an AI rebuilt in minutes
A page rebuild is not a channel change. It is a change to every channel at once, landing on one date, and it is usually not written down anywhere.
- An AI rebuilt page breaks channel attribution
- Incremental sales, not correlations
- When everyone clones the same offer block
- Machine readable attribution for AI agents
The rule underneath
All twelve reduce to the same discipline. Name the comparison, or admit there is not one. Read the confidence interval rather than the midpoint. Report what could not be resolved instead of filling it in with a number that looks like the others.
That is what causal inference is for, and it is why last-click attribution keeps producing confident answers to questions it cannot see.
Where a read fits
For the wider set of attribution questions, see the attribution answers hub and the AI era answers hub.
Frequently asked questions
Why does reported ROAS stay flat while I scale?
Because the platform keeps claiming the easy conversions it already had while the newly reached audience converts worse. The average of a strong base and a weak margin can look unchanged while the marginal return falls.What makes an attribution finding defensible?
The comparison it is measured against is named, the confidence interval is published, the share of revenue covered is stated, the unresolved channels are listed rather than scored, and the condition that would reverse the conclusion is written down.Does rebuilding a product page affect channel attribution?
Yes. A page change moves site wide conversion rate on a single date, so every per channel comparison spanning that date is partly measuring the rebuild. Record site changes in the same changelog as campaign changes.
Go deeper: Incrementality testing, 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.
- Attribution SoftwareAttribution Software measures campaign impact by tracking customer interactions across touchpoints. It assigns value to each channel, showing what drives conversions.
- Causal AttributionCausal Attribution uses causal inference to determine which marketing touchpoints genuinely cause conversions, not just correlate with them.
- Causal InferenceCausal inference is the set of methods for estimating how much an outcome changes because of a cause, rather than alongside it. It works by comparing what happened with an estimate of what would have happened without the cause.
- Confidence IntervalConfidence Interval is a statistical range of values that likely contains the true value of a metric. In marketing analytics, it quantifies uncertainty around estimates, indicating the precision of an outcome or causal effect.
- CorrelationCorrelation is a statistical measure showing a relationship between variables; it does not imply causation.
- 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.
- ShopifyShopify is an ecommerce platform for creating online stores and selling products. Attribution modeling shows which marketing channels drive traffic and conversions within Shopify.