How to use Revenue Funnel Analysis
Conversion-count funnels reward the cheapest buyers. Revenue funnel analysis weights each stage by AOV and margin, surfacing the cohorts that actually move the P&L.
Revenue Funnel Analysis
A funnel view that weights each stage by revenue (and ideally margin), not just conversion count, so the cohort that pays — not just the cohort that converts — drives optimization decisions.
Revenue funnel analysis is a way of reading your e-commerce funnel that replaces stage counts with stage value. Instead of asking "what percentage of sessions reach checkout?", it asks "what share of potential revenue makes it to checkout?" — which is the same thing only when every visitor is worth the same amount. They never are.
In practice it pairs stage-level conversion rate with average order value (AOV), and ideally margin, for each cohort you care about: traffic source, device, landing page, country, returning vs new. The output is a funnel where some segments are wider in revenue than they look in count — and others, usually the ones eating your CRO roadmap, are narrower.
Most funnel dashboards default to a count view because that's what GA4 and the analytics tutorials show you. You see 100,000 sessions narrowing to 2,400 purchases and you optimize toward whichever step has the worst drop-off rate.
The problem: a 3% conversion rate at €45 AOV and a 1.8% rate at €120 AOV produce wildly different revenue per visitor. Treating them as the same funnel — or worse, prioritizing the first because its conversion looks better — leaves money on the table every week the dashboard is open.
Why count-based funnels mislead you
A count funnel implicitly assumes every visitor carries the same expected revenue. The math behind "focus on the biggest drop-off" only holds if you're indifferent between losing a €30 buyer and losing a €300 buyer at that step. You are not.
Take a typical apparel store. Meta retargeting traffic converts at 4.2% but averages €58 per order. Google branded search converts at 2.6% but averages €142. The count view tells you to double down on Meta. The revenue view tells you Google branded is contributing more cash per session and any leak there costs more to ignore.
The same distortion shows up between devices, between new and returning buyers, and between categories. Mobile usually converts worse but pulls more impulse purchases at lower AOV; desktop converts better on higher-consideration baskets. If you only see counts, you'll keep redesigning mobile checkout while desktop quietly underperforms on the orders that pay the rent.
The conversion-rate trap
Optimizing for site-wide conversion rate without weighting by AOV almost always shifts your mix toward cheaper orders. You'll hit your CR target and miss your revenue target in the same quarter.
Building the revenue-weighted view
Start with the same four or five funnel stages you already track — session, product view, add-to-cart, checkout start, purchase. For each stage and each segment, compute two numbers: visitors who reached the stage, and the AOV of those who eventually purchased from that cohort.
The stage-level revenue is then visitors × stage conversion rate × cohort AOV. Plotting that side-by-side with the count funnel is where the insight lives — the two shapes will diverge, and the divergence points to where revenue funnel analysis changes your roadmap.
Same funnel, two views: count share vs revenue share by traffic source
Share of purchases (count)
Share of revenue (€)
Meta retargeting drives 38% of purchases but only 24% of revenue — high volume, low basket. Google branded flips it: 22% of purchases, 31% of revenue. Email sits between because returning buyers basket-build. The revenue lens reorders your priority list within thirty seconds.
Reading the funnel at the segment level
Once the revenue view exists, the next move is to compare stage conversion rate against revenue-per-session (RPS) for every meaningful cohort. RPS is the cleanest single number — it collapses CR and AOV into the only metric that closes against P&L.
A useful pattern: rank cohorts by RPS, then look at where in the funnel each cohort leaks hardest. High-RPS cohorts with a single weak stage are your best CRO targets — the upside per fixed percentage point is enormous. Low-RPS cohorts with broad weakness are usually a traffic-quality problem, not a UX problem.
Funnel stage performance by segment — count CR vs revenue per session
| Segment | Add-to-cart rate | Checkout completion | AOV (€) | Revenue per session (€) |
|---|---|---|---|---|
| Mobile, new visitor, paid social | 9.2% | 58% | 52 | 0.94 |
| Mobile, returning, email | 14.1% | 71% | 78 | 2.43 |
| Desktop, new, organic search | 7.8% | 66% | 118 | 2.18 |
| Desktop, returning, direct | 12.4% | 78% | 164 | 4.71 |
| Tablet, new, paid search | 8.6% | 62% | 96 | 1.74 |
Notice the desktop-returning-direct row. The add-to-cart rate isn't the highest on the page, but the AOV and stage completion combine into nearly five euros of revenue per session — five times the mobile-new-paid-social cohort. A one-point improvement here is worth roughly five points there. Count-based prioritization would never surface that.
Turning the analysis into roadmap decisions
Revenue funnel analysis is upstream of test selection, not a substitute for it. Once you've ranked stage-cohort leaks by revenue exposure (visitors × current leak × cohort AOV), you have a euro-denominated backlog. The top three or four items typically account for 60-70% of total addressable upside — that's where your next sprint goes.
Pair it with broader funnel optimization work: form-field tests, PDP layouts, shipping-threshold messaging. The revenue lens doesn't change the tactics — it changes which page, which device, which cohort you run them on first. That ordering decision is usually worth more than any individual test win.
A simple weekly habit
Pull RPS by traffic source and by device every Monday. When a source's RPS drops more than 15% week-on-week without a corresponding CR drop, the leak is in basket value — usually a promo, inventory, or recommendation issue, not a UX one.
Revenue funnel analysis FAQ
Funnel optimization is the broader practice of improving stage-to-stage conversion. Revenue funnel analysis is one lens inside it that weights each stage by the money flowing through, so the optimization order reflects euro impact rather than count drop-off. You use it to choose what to optimize; you still use the rest of the funnel optimization toolkit to actually move the number.
GA4 can get you most of the way with custom explorations that include purchase revenue and a session-scoped funnel, though pivoting AOV by cohort is clunky and the sampling at higher cardinalities hurts. Most teams build a Looker Studio or BI layer on top, or use a dedicated analytics tool that natively reports revenue per stage.
You want at least 200-300 purchases per cohort per month before AOV stabilizes enough to trust. Below that, segment by fewer dimensions — e.g. device only, or source only — instead of cross-cutting. Noise in AOV is the main risk; a single €800 outlier order can flip cohort rankings if your sample is thin.
Margin if you can get it cleanly, AOV otherwise. Margin matters most when product categories have very different gross margins — beauty consumables versus electronics, for example. If your catalogue is fairly uniform, AOV is a good-enough proxy and is far easier to keep current.
Revenue funnel analysis sits one layer below attribution. Attribution decides which channel gets credit for a purchase; the revenue funnel then tells you where within that channel's journey the value is concentrating or leaking. Use whatever attribution model your team has agreed on — the funnel analysis is robust to the choice.
Revenue per session (RPS) is total revenue divided by total sessions for a cohort. It collapses conversion rate and AOV into a single number that maps directly to P&L. Optimizing RPS rather than CR avoids the trap of improving conversion by shifting mix toward cheaper orders.
Yes, and for most DTC stores you should. The primary metric for checkout, PDP, and pricing tests is usually revenue per visitor; conversion rate becomes a secondary guardrail. The trade-off is statistical: revenue has higher variance than a binary conversion, so tests need slightly more traffic or more time.
It works, but you swap AOV for first-order value plus expected LTV contribution per cohort. The principle is identical — weight each funnel stage by the value flowing through it — but the value definition extends beyond a single purchase. For pure subscription, weight by first-month MRR plus a cohort-level retention factor.
Look at top-line RPS by source and device weekly, full segment breakdowns monthly, and rebuild your prioritized leak ranking once a quarter or after any significant traffic-mix change. AOV drifts more slowly than CR, so cohort rankings tend to be stable enough for quarterly cadence on the deep view.
Cutting low-RPS cohorts entirely. Low RPS doesn't mean worthless — it often means top-of-funnel awareness traffic that converts later, on a different device, in a different cohort. Use revenue funnel analysis to prioritize optimization, not to prune channels. Channel decisions need the full attribution picture.
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.