> ## Content Index > Fetch the complete content index at: https://conversion.club/llms.txt > Use this file to discover other available public pages before exploring further. # The Prism: A Framework That Takes CRO Beyond the Funnel - URL: https://conversion.club/the-prism-a-framework-that-takes-cro-beyond-the-funnel/ - Published: 2026-06-16T09:03:56.000Z - Updated: 2026-10-19T20:26:18.000Z - Author: Remi Migliasso - Tags: CRO, Guides Most CRO programs I've audited suffer from the same problem: they optimize the journey, and only the journey. Product page, funnel, checkout. That's useful, but it's not enough. After putting so many e-commerce brands under the microscope, I ended up mapping the real drivers that decide *contribution margin per visitor* (CMV). There are five main families, and the journey is only one of them. ## It took me a while to understand this At first, I did CRO like everyone else: I'd land on a site, analyze the funnel, hunt for friction, and propose tests. That's the job as it's taught, as it's sold, and as most agencies practice it. > And it works... partially. The problem is that this approach asks a very narrow implicit question: **"How do we get more of the visitors who land on this site to convert?"** It treats traffic quality, offer structure, post-purchase behavior, and geographic mix as fixed. It optimizes within a given perimeter and never challenges that perimeter. That's [funnel optimization](https://conversion.club/how-to-optimize-an-e-commerce-product-page-the-complete-audit/), and it's only one part of CRO. The distinction matters because it has direct consequences for the results you can expect and the questions you choose to ask. I started thinking differently when I realized that some of the tests I'd run showed a statistically positive result, but reality told a different story. The deeper you dig, the quicker you see how much more complex it all is. **Something was slipping past me.** I was using tools without understanding what they were technically spitting out. I don't like that situation: using a tool without knowing how it works. --- *I remember what a friend told me: in his university program, the algorithm professors made students write out the mechanics on paper before touching a machine.* ![Visual "cdr, cprdr, cmprndre, comprendre!": write out the mechanics before using the machine](https://storage.ghost.io/c/b5/a7/b5a79c8d-ed77-4459-9729-b694bd04720c/content/images/2026/10/8483188c-3a61-49e2-88ad-7497f5c109fc-1.webp) cdr, cprdr, cmprndre, comprendre! --- That's when I realized I'd never be comfortable selling CRO without understanding it. And that CRO, well... is simply probability. I've been taking stats courses for a while now (you can start here if you're curious: [Introduction à la statistique avec R](https://www.fun-mooc.fr/fr/cours/introduction-a-la-statistique-avec-r/?ref=conversion.club).) As I kept going, listening to interviews and discovering Simpson's paradox and other side effects, I naturally came to see that CRO isn't about conversion but about margin. --- > The metric that really matters isn't conversion rate. --- It's **contribution margin per visitor**: what a visitor actually earns the brand, after variable costs. Dozens of drivers influence this metric that conversion rate can't see: the quality of incoming traffic, the structure of the offer, repurchase frequency, the ability to convert in non-domestic markets. After working on these drivers for long enough, I grouped them into five families and [defined CRO](https://conversion.club/the-c-in-cro-comprehend-1-3/) as a **Prism**. --- ## The Prism: five drivers, one north star metric The Prism isn't a checklist. It's a lens that lets you ask, for any e-commerce brand: which of these five drivers will give the best return on the next euro invested in CRO? **The north star metric stays constant: contribution margin per visitor.** What changes from one brand to the next is the driver with the highest potential impact at a given moment. Because a brand is a living thing, the driver changes from one period to the next. ![Diagram of the Prism: the five drivers around contribution margin per visitor](https://storage.ghost.io/c/b5/a7/b5a79c8d-ed77-4459-9729-b694bd04720c/content/images/2026/10/image--1-.png) --- ### Driver 1: Acquisition CRO starts before the site, because the quality of incoming traffic largely determines what can happen next. **A visitor who arrives with no intent to buy won't convert, no matter how good the product page is.** A channel that sends poorly qualified visitors will drag down every downstream conversion ratio, and no one will understand why the A/B tests aren't "taking." ****What I look at on this driver:** ****Blended CAC**, which I always read alongside CLTV. One without the other tells you nothing. A high CAC can be perfectly healthy if customer lifetime value justifies it. This cross-reading of Acquisition and Loyalty gives a holistic view of how profitable your traffic really is. The **Acquisition** diagnostic is often the most profitable one to run on brands with significant paid activity. > Trade-offs are fast and the gains are direct. ### Driver 2: Journey This is CRO's historic territory and it deserves continued attention, but with one important caveat. A journey has two types of problems: - **Friction**: something gets in the way of a purchase that's already been decided. - **Intent blockers**: something keeps a decision from forming. Most CRO programs only address friction because it's easier to spot and easier to test, but the impact is structurally limited. **Converting intent means tipping the balance for someone who's on the fence.** That rarely happens at checkout. It happens on the product page, in how the value is argued, in the reassurance elements that answer real objections. !["The Line of Desire," photograph by Jamie Hawkesworth](https://storage.ghost.io/c/b5/a7/b5a79c8d-ed77-4459-9729-b694bd04720c/content/images/2026/10/8fdd9c17-864b-4b16-a0e1-cbb692ea23f3.webp) [The Line of Desire](https://fr.wikipedia.org/wiki/Ligne%5Fde%5Fd%C3%A9sir?ref=conversion.club) @Jamie Hawkesworth ****What I look at on this driver:** Drop-off rates by step and by segment, the points where intent fades, the quality of the product argument. ### Driver 3: Offer This is the most underused driver, and probably the most powerful. The offer is what the brand sells, how it presents it, and at what price. It isn't strictly CRO in the "classic" sense, and that's exactly why no one touches it. --- > **Too often, we optimize the container without ever questioning the contents.** --- Yet a hypothesis about pricing structure, range hierarchy, or cross-sell opportunities can have ten times the impact of a button test. ****What I look at on this driver:** The catalog Pareto (which 20% of products generate 80% of the contribution?) and sell-through (which products move, which stall?). These two readings are often enough to reveal merchandising trade-offs that A/B tests will never catch. ### Driver 4: Loyalty Acquiring a customer is expensive. Losing them after their first purchase is a direct destruction of value. Yet nearly all CRO programs stop right there. What happens after the purchase is left to the CRM team, often disconnected from the overall optimization logic. Cumulative contribution margin by customer cohort is nonetheless one of the most telling metrics for judging the health of an e-commerce business. A brand that improves its repurchase rate by five points doesn't need to buy five more points of traffic to grow. That's capital-efficient growth. ****What I look at on this driver:** The natural repurchase rate, the LTV gap between loyal customers and one-time buyers, the critical moments in the post-purchase lifecycle, and how well CRM sequences address the real barriers to repurchase. ### Driver 5: Cross-Regional This is the Prism's most unusual driver and the only one that works differently from the other four. Cross-Regional isn't a sequential driver. It's a transverse dimension: if the diagnostic shows geography is a priority growth vector, it becomes a filter applied across the other four drivers. You then look at **Acquisition** by market, **Journey** by market, **Offer** by market, and **Loyalty** by market. In practice, the starting signal is often simple: a country that accounts for 15% of traffic but 4% of revenue. That's a gap versus the domestic market that nothing explains. From there, the barriers can be logistical (shipping, cost, delivery time), perceptual (trust, brand awareness), or experiential (currency, language, localized content). Identifying which one it is points you to the driver to work on first. ****What I look at on this driver:** Conversion rates and contribution by geographic zone, and the additional CM potential if those markets matched the domestic one. --- ## How to use it The Prism isn't meant to be worked through in order. It's a diagnostic and prioritization tool. The first step is to instrument each driver with data: - Contribution margin per visitor (North) - Blended CAC (Acquisition) - Pass-through rate by step and segment (Journey) - Product performance and basket structure (Offer) - Repurchase rate and LTV by cohort (Loyalty), - Performance by geographic zone (Cross-Regional) This instrumentation almost always reveals one or two drivers with a disproportionate return potential, and that's where the real work begins. Not by testing everywhere, but by concentrating effort where the potential delta in contribution margin is highest. The question I ask on every account: **which driver gives our next invested euro the highest marginal return?** The answer changes with the brand's context, its growth stage, and what's already been worked on. Two brands with the same contribution margin per visitor can have radically different investment programs. **That's the whole point of the Prism: it forces you to look at the full picture before deciding where to dig.**