Revenue per visitor, the metric that ends the traffic argument

Two channels. One sent 8,000 visitors, the other sent 2,000. Which one deserves next month’s budget?

Traffic says the first. It is louder and the chart is bigger. But the question you are actually asking is which visitor is worth more, and there is a single number that answers it.

The calculation

Revenue in a period, divided by visitors in the same period.

revenue per visitor = revenue ÷ visitors

That is it. A month with $4,000 in revenue and 8,000 visitors is $0.50 per visitor. A month with $4,000 and 2,000 visitors is $2.00. Same money, four times the quality of attention.

Once both channels have a number, the ranking stops being an argument about which chart looks better and becomes arithmetic.

Why it beats conversion rate on its own

Conversion rate tells you how many people did the thing. Revenue per visitor tells you what the average visit was worth, which folds in order size as well as frequency.

The difference matters when a change moves the two in opposite directions. Drop your price and conversion rate climbs — more people buy. But if the price dropped further than the volume rose, revenue per visitor falls, and the “successful” experiment lost money. Only one of those two metrics would have caught it.

The reverse happens too. An enterprise tier converts terribly and looks like a failure by conversion rate, while quietly lifting revenue per visitor because the few who convert are worth ten of the others.

Three ways people get it wrong

Mismatched periods. Revenue for the month divided by visitors for the week is a meaningless number, and it is the most common mistake — usually because the two figures came from two tools with two different default ranges. Fix the range on both before dividing. This is the single strongest argument for having revenue and traffic in the same dashboard: you cannot accidentally compare August to last Tuesday when both series sit on one chart.

Mixed currencies. Adding €10 to $10 to get 20 of something is not a metric. Read one currency at a time, or convert deliberately with a stated rate and date convention.

Treating it as attribution. Site-wide revenue per visitor is a fine health metric. Per-channel revenue per visitor requires you to know which revenue came from which channel, and dividing total revenue by one channel’s visitors does not do that. Either fire a conversion event you can attribute, or accept the number as directional and stop making six-figure decisions with it.

Lag, and how to handle it

Revenue per visitor assumes the visit and the purchase happen close enough together to sit in the same period. For an impulse purchase that is fine. For a product with a 30-day consideration cycle, this week’s revenue was earned by traffic from a month ago, and dividing them produces noise.

Two adjustments help. Widen the window — monthly instead of weekly smooths most of it out. And compare like-for-like periods rather than reading the absolute number: this month against last month tells you something even when the level itself is distorted by lag.

Reading it in practice

Look at the trend, not the value. There is no benchmark for revenue per visitor because it depends entirely on your price point; a $2,000 B2B tool and a $9 ebook cannot be compared and neither can be compared to yours. Your own last quarter is the only benchmark worth having.

Then use it as a filter for the traffic chart. When traffic climbs and revenue per visitor drops, you found new people who do not want what you sell — worth knowing before you buy more of them. When traffic is flat and revenue per visitor climbs, something on the product or pricing side is working, and it deserves the credit that a traffic-only dashboard would have handed to marketing.

In sonex, both halves of the division sit on the same Overview chart: revenue as a focusable series on its own money axis, visitors beside it, over whatever range you pick. Connecting Stripe or Polar takes a minute and is included on every plan (Revenue). The arithmetic is still yours to do — but at least the two numbers will be describing the same period.