CBSplit CMS Wiki

This page is read only. You can view the source, but not change it. Ask your administrator if you think this is wrong.
====== Why two ClickBank offers with the same EPC behave differently ======

===== The EPC assumption =====

EPC is often treated as a shortcut to truth.

If two offers show the same EPC, they are assumed to be equally good.
Equally profitable.
Equally scalable.

This assumption breaks quickly in real funnels.

CBSplit was built to explain why identical EPC does not mean identical behavior.

===== EPC is an average, not a guarantee =====

EPC is a blended metric.

It averages revenue across:  
  * Different traffic sources
  * Different buyer intent levels
  * Different funnel paths
  * Different payment outcomes

Two offers can arrive at the same EPC through completely different mechanisms.

CBSplit looks beneath the average.

===== One EPC can hide clean revenue, the other fragile revenue =====

Offer A may generate EPC through:  
  * Clean approvals
  * Low retry dependence
  * Minimal refunds
  * Stable subscriptions

Offer B may generate the same EPC through:  
  * Heavy retries
  * High refund rates
  * Aggressive upsells
  * Early churn

EPC alone cannot distinguish these.

CBSplit can.

===== Payment behavior shapes offer durability =====

Behind identical EPC numbers, payment behavior often differs.

Key differences include:  
  * Decline frequency
  * Retry success rate
  * Fallback payment usage
  * Processor friction

Offers that rely on retries behave differently at scale.

CBSplit tracks payment behavior explicitly.

===== Refund timing distorts EPC =====

Refunds do not impact EPC immediately.

They arrive:  
  * Days later
  * After fulfillment
  * After customer dissatisfaction

Two offers may show identical EPC today.

One may collapse tomorrow under refunds.

CBSplit measures revenue after refunds, not before them.

===== Upsell paths change revenue quality =====

Offers with similar EPC can differ in:  
  * Upsell acceptance rates
  * Downsell reliance
  * Subscription attachment
  * Post-purchase intent

These differences shape long-term value.

EPC compresses them into a single number.

CBSplit preserves the structure.

===== Traffic interaction matters =====

Offers do not exist in isolation.

Their behavior depends on:  
  * Traffic source quality
  * Messaging alignment
  * Funnel expectations
  * Audience sophistication

Two offers with the same EPC can react very differently to the same traffic.

CBSplit observes these interactions over time.

===== EPC ignores operational stress =====

At scale, operational stress reveals truth.

Offers differ in how they handle:  
  * Volume spikes
  * Payment outages
  * Support load
  * Processor scrutiny

EPC does not capture stress response.

CBSplit does.