Revenue Per Visitor Calculator

This calculator exists to catch one specific mistake: shipping a variation because conversion went up, without noticing it sells cheaper. It is different from its siblings. The conversion rate calculator measures what percentage buys and compares it to your sector. The conversion rate impact calculator projects what a rate gain becomes in money, assuming order value stays flat. The funnel calculator shows where people vanish. Here the subject is the one metric that folds volume and value into a single number, and what happens when those two parts move in opposite directions.

How to use it

1. Fill in visitors, orders and total revenue for version A (the control), all from the same period and on the same base: either sessions or unique visitors, never mixed.
2. Do the same for version B (the variation). If you only want your current site RPV, fill in version A alone and read the first box.
3. Enter the monthly visitors of that flow so the tool can project the annual impact of the difference.
4. Read the verdict: it tells you who is ahead in money per visitor, not in orders.
5. Look at the second table. It shows how much of the RPV difference came from conversion and how much came from average order value. When the two signs disagree, the red warning appears.

How it works: the formula

The main calculation is a division. The interesting part is the factorization, which is why RPV never lies about a trade between volume and value:

The three terms add up to the RPV difference exactly, with no residue: it is the algebraic factorization of a product whose two factors both changed. The interaction term is small when both changes are small, and it is what stops the decomposition from counting the same money twice. In practice you read the first two numbers: if their signs disagree, the variation made a trade, and RPV is the judge of that trade.

Worked example (reproduces the default result)

With the numbers already filled in, a real discount test. Version A: 20,000 visitors, 500 orders, $120,000 in revenue. Version B (with a 12% coupon in the banner): 20,000 visitors, 560 orders, $117,600 in revenue.

• Version A: conversion = 500 ÷ 20,000 = 2.50%; order value = 120,000 ÷ 500 = $240.00; RPV = 120,000 ÷ 20,000 = $6.00.
• Version B: conversion = 560 ÷ 20,000 = 2.80%; order value = 117,600 ÷ 560 = $210.00; RPV = 117,600 ÷ 20,000 = $5.88.
• Conversion went up 12% (from 2.50% to 2.80%), order value fell 12.5% (from $240 to $210), and RPV fell 2.0% ($0.12 per visitor).
• Conversion effect: 0.003 × 240 = +$0.72. Order value effect: 0.025 × (−30) = −$0.75. Interaction: 0.003 × (−30) = −$0.09.
• Sum: 0.72 − 0.75 − 0.09 = −$0.12 per visitor. At 20,000 visitors per month, that is −$28,800 per year.

That is exactly what the tool shows when the page loads. Notice the size of the mistake: variation B looks like a clean 12% conversion win, a number that goes up on any dashboard and earns applause. In money, it costs almost thirty thousand dollars a year. The coupon bought orders with someone else's margin.

How to read it, and where it misleads

RPV is an average, and an average of money is fragile. Order value distributions are skewed, with a long tail of large purchases: one corporate buyer placing an order ten times the normal size can flip the winner of the test on its own. Before trusting an RPV difference, check whether it survives removing the largest order from each group. If it does not, you have an outlier, not a result.

That fragility has a practical consequence: the RPV difference you see here is descriptive arithmetic, not statistical inference. A two proportion z-test, which is what powers the significance calculator, applies to conversion rate, not to revenue. For revenue, the honest route is to confirm the conversion part with the proportions test and treat the order value difference as a signal to investigate, not as a verdict. If your volume allows it, a Bayesian test on revenue or a bootstrap over the order value distribution is the right instrument.

Three measurement traps also wreck the comparison before statistics get a chance. A different base between versions (one in sessions, one in users) mechanically inflates one side. Different periods drag seasonality and calendar promotions into the test. And revenue recorded with tax, shipping or cancelled orders in one version but not the other creates a difference where none exists. Standardize the definition of revenue before reading the number.

From RPV to the decision

RPV is for choosing between versions, and to do that it has to enter the test as the primary metric from the start, not as a consolation after conversion disappoints. Picking the metric once you have seen the result is like picking the finish line after the race.

• Declare the primary metric in the hypothesis. If the test touches price, discount, shipping, bundles or upsells, the primary is RPV, not conversion.
• Size the experiment before launch in the sample size calculator. Revenue tests need more volume than click tests, because the variance is larger.
• Track conversion rate and average order value as secondary metrics. They are the diagnosis: they tell you why RPV moved.
• Be suspicious of any variation that wins on conversion and loses on order value. It is almost always trading margin for volume, and the math only works if the customer comes back.

The full context on choosing metrics and prioritizing improvements is in the conversion rate optimization guide. If you do not know yet whether your conversion rate is good or bad, start with the conversion rate benchmarks.

FAQ

What is revenue per visitor (RPV) and how do I calculate it?

Revenue per visitor is total revenue divided by the number of visitors in the same period. If 20,000 visitors generated $120,000, RPV is $6.00. It is the same as multiplying conversion rate by average order value: 2.50% × $240 = $6.00. Both routes give the same number, and that identity is exactly what makes the metric useful.

Why use RPV instead of conversion rate in an A/B test?

Because conversion rate counts orders and RPV counts money. A variation with a coupon, a cheaper bundle or a low free shipping threshold makes more people buy and each of them spend less. Conversion goes up, the till goes down, and anyone watching conversion alone ships the wrong version. In ecommerce, where order value varies a lot, RPV is the more honest primary metric.

Are RPV and revenue per session the same thing?

It depends on the denominator you use. If it is sessions, the result is usually called revenue per session; if it is unique visitors, revenue per visitor. The arithmetic is identical, the number is not. What you cannot do is mix them: use the same base on both versions and over the same period, or the comparison means nothing.

Does a higher RPV mean the variation won?

No. RPV has far higher variance than a conversion rate, because a single huge order moves the whole group average. An RPV difference sizes the prize; it does not prove the prize exists. Confirm the conversion part in the significance calculator and look at the order value distribution before you decide.

What is a good RPV?

There is no universal number, because RPV depends on your price point. A furniture store with an RPV of $40 and an accessories store with an RPV of $2 can both be perfectly healthy. The correct use is comparing you against you: same page, same channel, different periods or versions. Cross industry comparison only misleads.

How do I project the annual impact of an RPV gain?

Multiply the RPV difference by the monthly visitors of that flow, then by twelve. A gain of $0.20 per visitor on 20,000 monthly visitors is worth $4,000 per month and $48,000 per year. It is a linear projection: it assumes stable traffic and channel mix and ignores seasonality.

Keep going

With RPV in hand, the next step is testing properly. Start with the CRO guide, size the experiment in the sample size calculator and validate the conversion part in the significance calculator. If you do not run tests yet, start with what A/B testing is.