How to use revenue per session to find your next fix
Most small sites count visits and orders, then argue about which page to change next. Revenue per session ends the argument: it puts a value on every visit, so the page, source or device that is not paying for itself becomes visible.
Short answer: revenue per session is your total revenue divided by your total sessions — what one visit is worth to your business on average. It is the ratio that connects traffic to money, and it turns a vague question ("is this page working?") into a decision: fix the page with the lowest value per visit first.
What revenue per session is (and what it is not)
Revenue per session (RPS) is one division: total revenue divided by total sessions, over the same period, for the same site. If a month brings €2,000 in orders and 4,000 sessions, a visit is worth €0.50. That example is illustrative; the only numbers that matter are yours.
It sounds trivial, and yet most small sites never compute it. What they watch instead is the pair everyone knows: how many visits came, and how many of them converted. Neither number knows anything about money. A page can double its traffic and cost you profit; a source can bring a tenth of the visits and pay for the whole month.
| Metric | The question it answers | What it hides |
|---|---|---|
| Revenue per session | What is one visit worth on average? | Nothing about volume or margin; one unusually large order can lift a small sample |
| Average order value | How much does a buyer spend? | Everyone who did not buy. Their visits still cost you content, design and hosting |
| Conversion rate | How often does a visit turn into an order? | The size of those orders. Many small orders and one large one look identical |
| Revenue | How much did the month bring in? | Whether the growth came from more visits or from richer visits — two different problems |
| Sessions | How much attention arrived? | Value. A doubling of traffic is a loss if the value per visit halves |
The five readings are not rivals; they are one story with five chapters. Revenue per session is simply the chapter that tells you whether the traffic you worked to get is worth what it costs you to get it.
One identity is worth keeping in your head, because it splits every conversion problem in two: revenue per session = conversion rate × average order value. If that number is weak, either too few visitors buy, or the ones who buy spend too little. Those are two different diagnoses, and they lead to two different fixes — which is why I look at this pair together and never at the conversion rate alone. The same discipline applies when you read a departure: bounce rate and exit rate look almost identical and point to different leaks.
How to calculate it, step by step
Four decisions, then it is arithmetic. The decisions matter more than the division, because they are what makes two months comparable.
- Fix the period and the scope. One site, one period, both figures from the same window. Use the same time zone in your shop backend and in your analytics, otherwise a busy evening at the end of the month lands in different weeks in the two reports.
- Decide what counts as revenue. Completed payments only, net of refunds and cancellations, with tax treated the same way every month. If you sell subscriptions, count the month you actually collected rather than the full contract value. Write the rule down in one sentence — a definition that lives in someone's head changes silently.
- Count sessions, not users. Analytics reports both. Use sessions, because sessions are what your pages and campaigns produce, and they match the way you buy traffic. Users smooth over repeat visits and quietly change the meaning of the number.
- Compute it for every view you will compare. One figure for the whole site is a starting point, not an answer. Add a row per source, per landing page and per device — same formula, narrower scope.
No shop? Assign an estimated value to a lead: average deal value multiplied by your close rate, kept identical over time. Label it as an estimate. Comparisons between segments stay useful even when the absolute value is approximate — that is how many B2B teams keep using this number without an e-commerce backend.
Where the raw material comes from is deliberately boring: revenue from your shop, invoicing tool or payment provider; sessions from your analytics. Two exports, one spreadsheet, ten minutes a week. What matters is doing it the same way every time.
Five ways to use it on a small site
Once the number exists, it starts answering questions that used to be opinions.
| Question | What to compare | What a gap usually means | Next step |
|---|---|---|---|
| Which channel is worth the effort? | Revenue and sessions per source | A source bringing volume with little money is feeding browsing, not buying | Read the intent behind the campaign before cutting it |
| Which page needs work? | Value per visit per landing page | A page that gets traffic and pays poorly is where the leak sits | Open the page, then watch the sessions that left it |
| Is mobile costing us? | Revenue per session per device | Mobile visitors arriving in numbers and buying far less than desktop ones | Check the first screen and the form on a real phone |
| Did the change help? | The same page, before and after | A flat line after a redesign is information: the problem was elsewhere | Keep or revert, then move to the next page on the list |
| Where do I spend the next hour? | Lowest value per visit with enough traffic | Small pages with thirty visits a month cannot be judged at all | Set a traffic floor and ignore everything below it |
Two habits make that table work. First, a traffic floor: I would not rank a page or a source with fewer than a few hundred sessions in the period, because the average of a handful of visits says nothing. Second, a second signal before any change. When a page pays less than its siblings you have located the problem, not explained it. The explanation comes from what the visitors did, and that is what finding out why visitors leave is for.
When revenue per session lies to you
The average is honest; the sample often is not. Five ways I have seen this number mislead a small site, all of them cheap to guard against.
- A very small sample. One €3,000 order among twenty visits turns a bad page into a star for a month. Wait for a traffic floor, and compare ranges rather than single figures.
- A change in the mix. A new campaign brings cheaper, earlier-stage visitors. Revenue grows, value per visit falls. Read the ratio alone and you will "fix" a campaign that is actually working.
- Money that comes back. Refunds, chargebacks and cancelled invoices belong on the revenue side of the division, or the number flatters you permanently. Reconcile with real receipts once a month.
- Two tools counting different visits. A blocked script, a consent banner or a cookie-less mode on one side and a full shop backend on the other give you two populations. Watch the trend of each figure, and treat the ratio as a direction rather than a decimal.
- Seasonality. A gifting month and a quiet month do not share a benchmark. Compare the same month last year, or compare segments inside the same period.
And one honest limit: revenue per session is an average, so it never describes an individual visit. The visitors who buy nothing are real people with a reason, and no ratio will tell you what it was. That is the boundary between measuring and watching — the same boundary that separates what a reach metric can prove from what only a session replay can show you.
What to put on the dashboard
You do not need a wall of charts. Three numbers, plus two splits, are enough to run a small site.
- Revenue per session, this week against your rolling four-week average. The gap is the story.
- Sessions and revenue, weekly. Together they explain why the ratio moved: more traffic, more money, or both.
- Two splits: value per visit by source, and by device. These decide where the next hour goes.
Review it in the same slot every week — same weekday, same ten minutes — so the comparison is like for like. Then act monthly rather than daily: rankings at source and page level need enough traffic, and a change you ship needs weeks to show up against the noise. Set one goal on the action that carries the money, so every later reading has something to be judged against — goals that answer a real question, not tallies of clicks.
On tooling, the honest position: revenue lives in your shop, sessions live in your analytics, and the ratio is a join you do yourself. SessionInsight covers the analytics half of that join for small sites — goals, sources, realtime, a cookie-less lightweight mode, email reports and shared dashboards — and keeps the sessions behind every drop-off so you can watch what happened instead of guessing. If you are still deciding whether you need a heavier suite at all, that is the right question to ask: choose your analytics by the question, not the feature list. You can see how it works or start on the free plan — one website, 50,000 pageviews a month, 30 days of history.
FAQ
What is a good revenue per session?
There is no universal good number: it depends on your prices, your margin and your model. The useful benchmark is your own history, plus the spread between your own sources, pages and devices. A channel at half the value per visit of another is the signal worth acting on, even when both look healthy.
Is revenue per session the same as average order value?
No. Average order value divides revenue by orders, so it only counts the people who bought. Revenue per session divides revenue by all sessions, buyers and non-buyers together. AOV tells you how much a customer spends; revenue per session tells you what a visit is worth to you.
How often should I look at revenue per session?
Once a week, in the same slot, on a rolling four-week view, is enough for most small sites. Daily readings are dominated by noise. Use the weekly number to notice a trend and the monthly view, split by source and device, to decide what to fix next.
Why did revenue per session drop while revenue grew?
Because traffic grew faster than revenue. A new campaign usually brings cheaper, earlier-stage visits that browse without buying, which lifts total revenue and lowers the average value per visit. That is not automatically a failure: it is a signal to look at the mix and at what those new visitors do before leaving.
Can I use revenue per session without an online shop?
Yes, with an estimated value per lead: average deal value multiplied by your close rate. Treat the result as an estimate, to compare segments with each other rather than as accounting. The comparisons stay useful even when the absolute number is approximate, as long as you keep the same estimate over time.
