How to use Revenue Funnel Analysis
Revenue funnel analysis weighs each cohort by what it pays, not just how often it converts — exposing the high-volume traffic that quietly loses money and the niche segments that fund the business.
Revenue Funnel Analysis
A funnel view that weights each cohort by revenue, not conversion count — combining conversion rate, AOV, and margin per segment.
Revenue funnel analysis is the practice of measuring your conversion funnel in euros rather than in checkouts. Instead of asking "which segment converts best?" you ask "which segment generates the most revenue per session, and at what margin?" The output is a per-cohort view that combines three numbers: conversion rate, average order value, and (where available) gross margin.
It sits inside the broader discipline of funnel optimization, but flips the optimisation target. A high-converting cohort that buys €25 items can be less valuable than a low-converting cohort that buys €180 bundles. Revenue funnel analysis surfaces that asymmetry so you stop optimising the wrong number.
Most funnel dashboards default to conversion rate because it's the easiest number to render. Add a percentage column, sort descending, ship. The problem is that conversion rate is a vanity metric in isolation — a 6% segment buying €30 add-ons is worth less than a 2% segment buying €250 statement pieces.
Revenue funnel analysis fixes the lens. You keep the funnel structure (landing → product → cart → checkout → purchase) and you keep the cohort splits (channel, device, audience, landing page), but every stage gets a revenue figure attached. Suddenly the cohort ranking changes, and so does the experiment roadmap.
Why a revenue lens beats a conversion lens
The core insight: conversion rate and AOV are usually negatively correlated within a store. Discount-led cohorts convert higher and pay less. Premium cohorts convert lower and pay more. If you only optimise the percentage, you systematically shift the mix toward your cheapest buyers.
On an apparel store with a €60 baseline AOV, a sitewide 10%-off banner often lifts conversion by 15-25% while compressing AOV by 8-12% and margin by more. The conversion dashboard celebrates. The P&L doesn't. Revenue funnel analysis would have caught the trade in week one.
The right scoreboard is revenue per visitor (RPV) per cohort, broken down by funnel stage. RPV = conversion rate × AOV, which means it captures both behaviours in one number. Watch RPV change and you can't trick yourself into shipping a test that only wins on volume.
Rule of thumb
If a test wins on conversion rate but loses on AOV, it's neutral at best until you've checked margin. If it wins on RPV with stable margin, ship it. If it wins on conversion rate and AOV, you've found a rare unicorn — replicate the pattern across the rest of the funnel.
Building the analysis
Start with the cohort dimensions that matter for your store. The usual short list: traffic source (paid social, paid search, organic, email, direct), device (mobile vs desktop), landing page type (PDP vs collection vs homepage), new vs returning, and — if you have it — first-touch campaign or audience segment.
For each cohort, pull four numbers across a recent 30-90 day window: sessions, orders, gross revenue, and discount value. From those you derive conversion rate, AOV, RPV, and discount rate. If you have GA4 with enhanced ecommerce wired correctly, all four are already there — they just aren't displayed together.
Conversion rate vs revenue per visitor by traffic source
Conversion rate (%)
RPV (€)
Notice paid search non-brand: it has the lowest conversion rate in the set, but its RPV beats paid social by 32%. A conversion-only view would tell you to cut non-brand spend. The revenue view tells you non-brand visitors buy bigger baskets when they convert — usually because they're researching specific high-intent products.
Reading the cohort table
Once you have RPV by cohort, sort the table by total revenue contribution (sessions × RPV), not by RPV alone. A cohort with €8 RPV but 400 sessions/month matters less than a cohort with €3 RPV and 40,000 sessions. The point of the analysis is to direct effort toward the cells where a 10% lift moves the P&L.
Then look for asymmetries: where conversion rate and AOV disagree sharply. Those are your most interesting cells. They're either underperforming on one axis (e.g. mobile converts fine but AOV is half of desktop — a checkout or product-density problem) or genuinely two different audiences sharing one funnel.
Typical RPV ranges by traffic source and vertical (DTC stores, €1-10M annual revenue)
| Traffic source | Apparel & accessories | Beauty & skincare | Home & lifestyle | Electronics |
|---|---|---|---|---|
| €3.80 - €6.20 | €4.50 - €7.80 | €3.20 - €5.40 | €2.90 - €5.10 | |
| Paid search (brand) | €3.50 - €5.50 | €4.20 - €7.00 | €3.00 - €5.00 | €2.40 - €4.30 |
| Direct | €2.80 - €4.60 | €3.60 - €6.00 | €2.50 - €4.20 | €2.20 - €3.80 |
| Organic | €2.20 - €3.80 | €2.80 - €4.80 | €1.90 - €3.30 | €1.60 - €2.90 |
| Paid search (non-brand) | €1.40 - €2.40 | €1.80 - €3.10 | €1.20 - €2.10 | €1.30 - €2.40 |
| Paid social (prospecting) | €0.90 - €1.80 | €1.20 - €2.30 | €0.80 - €1.60 | €0.70 - €1.40 |
Use ranges like these as a sanity check, not a target. If your paid social RPV is €0.40, you don't have a creative problem — you have a landing-page or product-fit problem, because the floor for the channel is roughly double that. If your email RPV is €9, your list is unusually engaged and the question is how to grow it without diluting it.
Turning the analysis into experiments
Every revenue funnel analysis should end in a ranked experiment list. For each cohort where RPV underperforms its expected range, write one hypothesis about what's compressing it: conversion side (friction, trust, speed) or AOV side (cross-sell, bundling, pricing presentation, free-shipping threshold).
The highest-leverage tests usually target AOV in already-converting cohorts. Lifting AOV by 8% in your email cohort might generate more incremental revenue than a 25% mobile conversion-rate lift, because the email cohort already converts at 6%+ and you're stacking the gain on a larger base.
Don't optimise the funnel into a single cohort
If you only chase whichever cohort has the best RPV today, you'll narrow your acquisition surface and cap growth. The goal is to lift RPV in the cohorts you can scale, not to abandon every cohort that isn't your best one. Paid social will never out-RPV email — but it's how you fill the email list.
Frequently asked questions
A standard funnel reports how many users move from stage to stage. A revenue funnel attaches a euro figure to each stage and each cohort, combining conversion rate with AOV. The result is revenue per visitor (RPV), which is the metric that actually correlates with P&L.
RPV = conversion rate × AOV. AOV measures what a buyer spends; RPV measures what a visitor is worth. AOV alone ignores the cost of unconverted traffic, which is why RPV is the better scoreboard when you're allocating paid-media budget.
Gross revenue is fine for diagnosis. For decisions involving paid spend or discount strategy, switch to gross-margin-per-visitor, because a discount-heavy cohort can have strong RPV but thin contribution margin. If you sell SKUs with very different margin profiles, include margin from day one.
30 days for fast-moving channels like paid social, 90 days for organic and email. Shorter windows are noisier; longer windows hide recent shifts. If you're past a major change — a site redesign, a price change, a new campaign — restart the clock from that date.
GA4 with the standard Shopify integration captures sessions, transactions, and revenue per source/medium, which is everything you need for a baseline RPV table. The friction is usually in cohorting — GA4's exploration reports are powerful but slow, and pivoting by two dimensions at once gets clunky.
Then RPV and conversion rate move together and the analysis tells you less. But check before you assume: AOV variance is almost always larger than people expect, especially across device (mobile typically 15-25% lower than desktop) and across discount-driven vs full-price cohorts.
Start with 5-8 cohorts on one dimension (usually traffic source). Once that's stable, layer device, then landing-page type. Anything beyond two dimensions at once becomes hard to read and the cells get too small for statistical confidence.
No — it sits inside funnel optimization as a reporting lens. You still measure stage-by-stage drop-off, run A/B tests, and fix UX friction. Revenue funnel analysis just changes which stages and which cohorts you prioritise.
Two options. Either restrict the analysis to first-order revenue (cleanest for acquisition decisions) or use 90-day customer value per visitor (better for LTV-driven channels like email). Mixing the two in one table makes the numbers incomparable.
Compute the ratio of (cohort RPV) ÷ (site-wide RPV) for each cohort. Anything under 0.6 with meaningful session volume is a problem cohort worth a hypothesis. Anything over 1.5 is a cohort to study and try to scale.
Get an AI expert review of your site
Paste your URL — Metricuno's AI runs the same heuristic checks a senior CRO consultant would, scoring your page and prioritising the fixes that'll move conversion fastest.