Good conversion rate: the five criteria that make yours comparable

A good conversion rate most often sits between 1% and 7%, depending on your situation and on what is counted: from 1.1% to 7.4% of unique visitors converted into a contact request in B2B, depending on the sector (First Page Sage, 2025); 6.6% of visitors as a median for a landing page (Unbounce, 2025).

This page is for the executive, or the marketing, e-commerce or product director, who wants to know whether their rate is good. It first gives the benchmarks by situation, then the five criteria that make yours comparable.

Benchmarks by situation

The table brings together the most useful published benchmarks. Each one is given with what the source divides by what: without that column, a benchmark lines up with no rate at all, yours included.

These rows do not compare with one another: some divide by sessions, others by unique visitors, visitors to a single page, paid clicks or an email list. A landing page median is not “better” than a new-visitor session rate: it counts something else, on another population.

Each cut has its own detailed table, listed in “Depending on your situation”. The original data are in the digital marketing statistics and, for paid clicks, in the SEA statistics.

What makes a rate good

A conversion rate is calculated in one way only: conversions divided by sessions, which is the GA4 “session conversion rate”. The formula and its variants are set out in what a conversion rate is.

A conversion is the action tied to revenue: an order, a qualified contact request. What counts as one on your site is decided once, starting from the definition of a conversion. Micro-conversions, such as an add to basket or a download, are tracked separately and never enter the rate.

With that definition set, a good conversion rate is recognised by three checks:

• It is profitable: it makes the acquisition cost compatible with the margin per conversion, either the basket multiplied by the margin in e-commerce, or the value of a lead multiplied by the closing rate in lead generation.
• It is improving: it beats your own history, calculated on a constant definition, and that trajectory counts for more than its level in any given month.
• It rests on sufficient volume: a rate calculated on a few dozen sessions and two or three conversions varies widely from one month to the next without anything having changed.

The arithmetic shows why profitability comes before the percentage. On 10,000 sessions, a rate of 3% with €20 of margin per order produces €6,000; a rate of 2% with €40 of margin produces €8,000. The second shows the lower rate and the better profitability.

Published averages and medians, those in the table included, give an order of magnitude. They set neither a threshold nor a target: they summarise other sites, on a denominator that is not necessarily yours.

What a conversion rate measures, and what it does not tell you about your performance

A conversion is what you want your visitor to do: a purchase for an online shop, a quote request for a software vendor. The rate relates that action to a reference population, two choices that belong to you: your rate is therefore not your neighbour's.

It measures what proportion of your audience did what you expected of them, over the period observed. It does not say:

• What the action brings in: a low-margin order weighs as much as a high-margin one.
• Whether the audience measured is the right one: a site that attracts off-target visitors shows a low rate without having done anything wrong.
• What it costs to bring the visitors in: yet the acquisition cost is what decides profitability.

Hence the five conditions without which a comparison does not hold, and which separate the rows of the benchmark table:

• The denominator: what you divide by what.
• The channel: where the visitors counted come from.
• The sales cycle: the time and the number of people it takes to decide.
• The offer: the price, the margin, the basket, the repeat business.
• The page type: what the page was supposed to obtain.

The denominator: the first reason why two rates do not compare

Two dashboards, the same site, the same month, two different rates: most often, it is the denominator that differs. Four units compete for that place, each one legitimate:

A disagreement between two dashboards most often comes from a different definition, sometimes from a measurement fault: a missing tag, an event counted twice, a view filter. Check both, then document once and for all the unit in the denominator, the actions counted and the set of pages covered.

That choice above all needs to be stable. If a change becomes unavoidable, it opens a new series: the old one is not recalculated, the two are kept side by side, with the switchover date noted. When you face an average found elsewhere, look for what it divides by what, over what period and on what volume.

The channel: a rate across all channels compares with nothing

Each channel, from organic to the answers of generative engines, brings in visitors who are not at the same point in their decision. An overall rate describes none of those audiences: hence one row per channel in the table, and three reading reflexes:

• Segment before commenting: the overall rate can fall while each channel improves, because the share of the least converting channel has grown.
• Expect from each channel what it can deliver: a visitor who came from a precise query has stated their need, unlike one picked up by a brand awareness campaign. The question is “is each channel doing what is asked of it, at the stage of the journey where it operates?”.
• Read the rate together with the volume: a rate that rises while sessions fall describes first of all a change in volume. This is common with entries from generative answers, which measurement tools identify poorly.

An excellent rate on a trickle of traffic is not worth an average rate on a volume that feeds the pipeline: the rate sheds light on volume, it does not replace it.

The sales cycle and the offer: why a low rate can be the right one

Between the click and the signature, an impulse purchase takes a few minutes; an enterprise contract takes weeks or quarters and several decision-makers. A long cycle does not on its own produce a low rate: measured on a booked meeting, a long-cycle offer often shows a high rate; measured on the signature, the same site looks very low.

Above all, a long cycle shifts conversions outside the period observed. Hence three measurements not to be confused:

On a long cycle, only the cohort says what a month's traffic has produced. And a rate that improves while lead quality deteriorates is bad news dressed up as good.

The offer then changes the real value of a conversion:

• The margin: selling far cheaper than the market drives the rate up, at the expense of the margin.
• The average basket: at an equal rate, a higher basket produces more value per session.
• Repeat business: two sites with the same rate are not in the same health if one sells once to customers who never come back, and the other several times a year to the same people.

A high rate is therefore not an objective in itself: what is steered is the value produced by the traffic. If your content or your offer leaves something to be desired, there is room for improvement whatever the benchmark for your sector: “is my rate good?” then gives way to “what does it bring in, and by how much can it improve?”.

The page type: a rate is judged page by page, not site by site

A site rate is a weighted average of pages that are not pursuing the same goal:

• The pricing page: the visitor arrives there set on comparing; highly converting, it is the least representative of the site.
• The product page: the intent is there, but the decision also turns on availability, delivery time or a competitor's price.
• The campaign landing page: one promise, one action, traffic selected upstream by the campaign.
• The article: the visitor was looking for information, not a supplier.

So a page is only compared with itself over time, and with its peers: at a minimum, one rate per page family, and a watch on how traffic is distributed between families.

This criterion also indicates where to look when a gap appears. If it is on a campaign page, examine the traffic and the promise; on a whole family, the page template. If it is across the entire site without any family moving, it is the mix that has changed, not your work.

Part of the gaps cannot be corrected from the content; knowing who they belong to saves rebuilding a page that is not at fault:

These levers are handled with their owners and on their timetable. The split is formalised once and for all: a list of pages, one owner per type of fix, a deadline. Without it, every gap in the rate reopens the same discussion.

Benchmark, history, segments: in that order

Once the five criteria are in place, compare your rate with three references, always in this order:

• Your history, on a constant definition: twelve months at a minimum, a full seasonal cycle, read on a rolling month-by-month basis; it is the only one whose method you know end to end.
• Your segments: channel, device, page type; a wide gap between two segments reveals potential more reliably than a published average.
• Published benchmarks, last: on an identical denominator, as an order of magnitude, never as a target.

When these readings contradict each other, history wins. Competitors are not on the list: their rates are not accessible, and a figure that cannot be checked does not serve as a benchmark.

Two rules protect this reading. Any change in the definition of the conversion or in tracking opens a new series, kept alongside the old one. And the decision window is set by the volume of conversions observed, not by the calendar; the reading cadence, for its part, belongs to tracking the conversion rate KPIs.

Then comes the target. The useful question is not “what is the median level?”, but “what rate can I reasonably aim for, above my history and while staying profitable?”. Three precautions so that the target holds up in front of a committee:

• set it by channel and by page type rather than for the whole site;
• date it, and attach the conditions that make it reachable;
• note, on the day you set it, the exact definition of the rate it applies to, denominator included.

Once the gap is identified, you have to decide which page to start with, what to change on it and how to measure the effect: that is the work of conversion rate optimisation.

Depending on your situation

Each situation has its reference article:

• You generate leads in B2B: the First Page Sage benchmark is read sector by sector in the average B2B conversion rate.
• You sell online: the order is your unit, with its own KPIs, in the e-commerce website conversion rate.
• You are judging a landing page: the Unbounce median and its variations are in the average landing page conversion rate.
• You are looking for your sector's benchmark: the data by activity are in the conversion rate by business sector.
• You steer a site across all channels: the benchmarks by channel are in the average website conversion rate.
• You are designing a campaign page: what makes its result vary is in the landing page conversion rate.
• You are looking for the headline indicator of a complete site: see the site conversion rate.
• Your entries come from organic search and generative engines: attribution is a matter of SEO conversion.
• You have identified the gap and need to prioritise: the decision window and the order of work belong to optimizing conversion rate.
• You only have an overall rate: the breakdown and the reading cadence are in the conversion rate KPIs.

To compare your rate by channel and by page type and build your own benchmarks, the performance reporting module brings these cuts together in a single dashboard.

Frequently asked questions (FAQ)

What is a good conversion rate?

A good conversion rate is a profitable rate that is improving: it makes the acquisition cost compatible with the margin per conversion, and it beats your own history on a constant definition. Published benchmarks place it without setting a threshold. To judge it, check the denominator, the channel, the sales cycle, the offer and the page type, then the volume it is calculated on.

What is a good conversion rate for a B2B website?

From 1.1% to 7.4% of unique visitors converted into a contact request, depending on the sector (First Page Sage, 2025): B2B software publishing sits at the bottom of the range, legal services at the top. This benchmark counts unique visitors, not sessions: it places your rate without serving as a target. Also track what becomes of the leads, through to closing.

What is a good conversion rate for an e-commerce website?

Count on 4.8% of paid clicks converted for e-commerce SEA (Odiens, 2025), and 2.4% conversion into a qualified lead or sale for retail and e-commerce (Ruler Analytics, 2026). The first measures paid traffic only, the second all tracked visits: they do not compare with each other. Your reference remains your own history of orders per session.

What is a good conversion rate for a landing page?

A landing page converts 6.6% of its visitors as a median, all industries combined (Unbounce, 2025). That level compares neither with that of a site page nor with that of an article: its traffic is selected upstream by the campaign. It varies with the intent of the source, the clarity of the promise and the friction of the form.

What rate should you aim for?

Aim for a rate above your history on a constant definition and profitable in light of your acquisition cost, never a published average. The level you already reach on your best pages and your best channels gives a documented, verifiable target. Set it by channel and by page type, date it, and record the definition of the rate it applies to.

How is the conversion rate calculated?

The conversion rate is calculated by dividing the number of conversions by the number of sessions, then multiplying by 100: 120 conversions for 4,000 sessions give 3%. Only actions tied to revenue count, such as an order or a qualified contact request; micro-conversions are tracked separately. Set this definition before any comparison.

Why is my conversion rate falling while my traffic is rising?

Most often, no channel has moved on its own: it is their proportion in the traffic that has changed. A rise in traffic dilutes the rate when the new visitors are less qualified, coming from broader queries, brand awareness campaigns or new information pages. Check the source, the intent of the queries and the rate by page and by device before concluding.

Can a lower rate be worth more than a higher one?

Yes, as soon as the value per conversion differs: a lower rate on high-margin orders can bring in more than a high rate on low-margin orders. So always compare the rate with what it produces: margin per conversion, acquisition cost, customer value over time. The percentage on its own does not say which site is in better health.