Site conversion rate: what your overall figure lets you decide

Measured per session, a site's conversion rate ranges from 1.7% for new visitors to 2.9% for returning visitors (Contentsquare, 2026): the overall figure on your dashboard depends first of all on what it aggregates. This page is for the executive or marketing director who tracks the site conversion rate and wants to know what decisions it supports.

It shows how to break the rate down and which pages to start with. The conditions that make two rates comparable belong to the criteria for a good conversion rate; the figures by channel are in the digital marketing statistics.

What the site conversion rate measures: scope, unit, calculation

The site conversion rate shows the share of sessions that end in a conversion: the revenue-linked action you have defined, whether a quote request, a qualified contact request or a purchase. Downloading a white paper or signing up to a newsletter are micro-conversions: they are tracked separately and do not count towards the rate.

What sets this rate apart from its neighbours is its scope: it takes the whole site as the unit of measurement, not a page, not a campaign, not a channel. That is what makes it the headline indicator for an executive committee, and also what makes it fragile. A merchant site, whose unit of conversion is the order, is read differently: that is the subject of the e-commerce website conversion rate.

The calculation is simple: a services site that records 20 conversions for 800 sessions over a month shows 20 divided by 800, i.e. 0.025, so 2.5% once multiplied by 100.

You still need to know what goes in the denominator. The rule is fixed: sessions, the "session conversion rate" in GA4, where visit and session mean the same thing. Depending on the unit chosen, the same month gives different rates:

• Sessions: the reference unit, the one your reports should show.
• Users: a person who came back three times counts once, which raises the rate without changing performance in any way.
• Clicks counted by an advertising platform: they measure the campaign and are read with it, not with the site.

The gap between these units is often larger than months of work would produce. Write the unit beside the figure along with the period chosen, and never change it again; a change of definition or tagging opens a new series, kept alongside the old one without recalculating it.

If what you are missing is the formula, or what "conversion rate" covers as against "transformation rate", the definition of the conversion rate settles both once and for all.

The site conversion rate is one of the most useful performance indicators, all the more so when the site has a commercial objective. It allows you to:

• bring consumption trends to light;
• go beyond the simple notion of page views or visits and obtain information on the profitability of your site;
• identify any malfunctions that stand in the way of its improvement;
• highlight the relevance of the strategies that have helped improve it;
• rationalise your marketing spend, spending sensibly for an optimal result.

A site rate is an average: what to break down before reading it

A site rate describes no visitor in particular. It is a weighted average: it adds together a pricing page that converts strongly and an article that barely converts at all, a mobile and a computer. The result is accurate, but has no operational reach until it has been broken down. In particular, it aggregates traffic sources that do not bring the same intent:

• organic search;
• paid traffic;
• direct and e-mail;
• recently, the answers of generative engines.

Analysed separately, these segments show which ones perform best and which can be improved. Without that breakdown, the most common effect is also the most misleading: the site rate falls while every channel improves. Nothing has deteriorated: the share of the least converting channel has grown in the mix. As long as you read the overall figure, you are commenting on the composition of your traffic while believing you are commenting on your performance.

Three breakdowns are enough to make a site rate readable, and they are set before the reading, not after. Each one points to a different decision:

Google Analytics 4 and steering platforms produce these cuts natively; Search Console, for its part, sees no conversion and only reports what comes before the click. The difficulty is deciding once which cuts are authoritative and sticking to them: a site rate is always commented on with its three cuts beside it.

What the rate does not say: value per visit, average order value, margin

The site conversion rate gains from being read alongside other indicators: raising the average order value appreciably can, in some cases, bring in more than optimising conversions. The rate on its own cannot settle that: it knows neither price, nor margin, nor what becomes of the customer afterwards.

The quantity that reconciles those readings is value per visit, rate and order value included: it alone can be compared with the cost of acquisition. The full economic criterion (order value × margin, or lead value × closing rate) is set out in the criteria for a good conversion rate cited above. The trade-off that follows is rarely stated:

• Few visitors convert: the work is on the rate.
• Those who convert buy little, or only once: the offer, the order value and loyalty enter the discussion.
• In both cases: gaining a point of rate improves revenue, since value per visit is proportional to the rate; what settles it is the quantified gain from each avenue, in euros and at comparable effort.

A worked example over two months of the same site shows why the rate on its own settles nothing:

Month B brings in more despite a lower rate. In the same way, if the rate stays flat while the volume of qualified visits and the order value rise, the site is doing what is asked of it. That is why a rate objective set on its own produces bad decisions: the fastest way to meet it is to reduce traffic.

Qualifying traffic: what the SERP promises, what the page delivers

Once the rate has been broken down, most of the gaps between entry pages come down to the distance between what the page promises in the search results and what it delivers once opened. This is the lever that belongs squarely to SEO, and the first to examine when the rate falls while traffic rises.

The check is done page by page, in three steps:

• Intent: what is the person typing that query really looking for: to understand, to compare, to choose or to buy?
• The promise: what do the title and description shown in the results, and then the first line of the page, announce to them?
• The offer: does the page make it possible to do, now, what that intent calls for?

A page that attracts an information intent and offers only a contact form will show a mediocre rate with nothing to reproach itself for; a page that attracts a purchase intent without a price, a condition or a way to get started will show a bad one for the opposite reason. What then gets fixed is almost never the layout, but three elements:

• Internal linking, directed towards the pages that actually carry the offer.
• Content that lifts objections instead of setting out generalities: comparisons and alternatives, terms and prices where they can be written down, frequently asked questions turned towards what is blocking.
• The title and description in the results, which decide the click-through rate and the quality of those who click: a narrower promise brings the volume down and the rate up, often a good exchange.

The same logic holds for the answers of generative engines. The page is no longer chosen by the reader but cited by the engine, and the visitor arrives with part of the answer already in hand. What has to be given to them is what comes next:

• the next step;
• the condition;
• the costing;
• the way to begin.

Prioritising: which pages to start with on a site that has thousands

The work is done page by page. Start at the crossing of qualified traffic and business potential: a heavily visited page with no commercial stake will bring in nothing, and a high-stakes page that nobody visits is an acquisition problem.

Three rules make that prioritisation workable on a site of thousands of URLs:

Each of these rules rests on an observation:

• The threshold: on a few dozen visits a month, one conversion more or less moves the rate by several points; what falls below it is handled by family.
• The list: two hundred pages do not make a list of priorities, whereas a few dozen tracked by name, reviewed at a fixed cadence, produce decisions.
• The order: a gap due to intent or to the promise is fixed quickly and measured well; one that comes from the offer calls for a trade-off outside your hands.

This list is also the basis for the discussion with the other teams: it names the pages, the gap observed and the type of fix expected.

The gaps that cannot be fixed from the content (funnel, usability, speed, tagging) are handled with their owners, following the division described in the guide to conversion rate optimisation. What you pass on to them is worth more than an opinion: the measurement by page and by device. A table of pages, an owner per fix, a deadline: the division is formalised once.

Traffic from AI engines: an entry point to track separately

AI answer engines now form a distinct entry point: the visitor who comes from them has already read part of the answer, and their behaviour on the site is measured rather than assumed. The GEO statistics track how this channel is developing. A site rate that moves deserves to be checked on that side, following three instructions:

• Segmenting AI traffic: to measure the real effect of GEO on the site conversion rate, isolate traffic coming from AI engines in your analytics tools and read its rate alongside that of organic traffic.
• Tracking cited entry pages: the pages most cited by AI become new entry points to your site. If they are designed for information but not for conversion, those visits leave with no next step.
• Taking attribution into account: a visitor may discover your brand through an AI answer, then come back directly or through a branded search. Classic attribution models then underestimate the contribution of GEO to the overall rate.

These instructions call for no new tooling, only the awareness that the attribution figure in your tool is a floor. Splitting a conversion between organic and generative answers is a matter of attribution, outside the site rate dashboard.

Placing yourself: your own history first, the sector benchmark second

The benchmark is built in a fixed order, from most to least reliable:

• Your history, at constant definition: twelve months minimum to cover a seasonal cycle, read on a rolling month-by-month basis.
• Your segments: channel, device and page type, with the same segmentation from one report to the next.
• Published benchmarks, last, with an identical denominator, as an order of magnitude and never as an objective.

How those published benchmarks are built, and their limits, are detailed in the average website conversion rate. An outside order of magnitude keeps its usefulness for opening the discussion in committee.

It is looked for by sector, and never across all sectors: that is what the conversion rate by business sector brings together, with one precaution: check what the source measures exactly, and on what population, before setting its value against yours.

Once that benchmark is in place, the opening question changes in nature. You are no longer trying to find out whether your rate is good: you know what it aggregates, you have named the pages to start with and what is not yours. Deciding what to change on those pages, in what order and how the effect will be measured is steering work, with a method of its own.

Continue reading

Depending on what you have to decide once your site rate is broken down, the reference article changes:

• You are setting up the breakdown: channel, entry page, device, campaign and the indicators coming from generative engines are detailed in the conversion rate KPIs.
• You have to attribute your conversions to the organic channel: saying what share belongs to you, and not only how many you make, follows the rules of SEO conversion.
• Your pages are prioritised: deciding what to change, in what order and how the effect is measured is the subject of optimising the conversion rate.
• You are judging a campaign page: its traffic is selected upstream and is read at another level, that of the landing page conversion rate.
• Your tool does not yet calculate the rate by entry page and by device: the setup and how to read the reports are described in Google Analytics.

To isolate traffic coming from AI engines and track the conversion rate of your entry pages, page by page, the performance reporting module makes it possible to segment AI traffic and measure its contribution to the conversion rate of your site.

Frequently asked questions

How do you calculate the site conversion rate?

The site conversion rate is calculated by dividing the number of conversions (sales, qualified contact requests) by the number of sessions over the period, then multiplying by 100. With 75 conversions for 3,000 sessions, it is 2.5%. The unit is written beside the figure and never changes; the rate is then broken down by channel, entry page and device.

What is a good conversion rate for a site?

A good site conversion rate is one that makes the cost of acquisition compatible with the margin per conversion and that improves on your own history. As an order of magnitude, 1.7% of new visitors' sessions and 2.9% of returning visitors' sessions convert (Contentsquare, 2026). These figures place your rate by type of visitor; they set neither a threshold nor an objective.

Which tools should be used to track the conversion rate?

Google Analytics 4 tracks conversions, journeys and attribution; Google Search Console reports what comes before the click (impressions, clicks, rankings) without seeing any conversion. Platforms such as Incremys combine SEO and GEO data in a single dashboard. What matters is consistency: one definition, one unit, the same cuts.

Which actions raise the conversion rate of a site?

The actions that are yours bring intent, promise and offer closer together: clarify the value proposition from the entry page onwards, align SEO and GEO content with search intent, and tighten the promise shown in the results. The other gaps (funnel, usability, speed) are fixed with the teams that own them, from the measurement by page and by device.

How do you improve the site conversion rate through SEO (and GEO)?

SEO improves the site conversion rate by bringing in visitors whose intent matches what the entry page offers: internal linking towards the pages that carry the offer, content that lifts objections, a more precise title and description in the results. In GEO, you isolate the pages cited by AI engines, track their rate and add the next step that visitor expects.

Which KPIs should be tracked alongside the conversion rate?

Alongside the site conversion rate, track the volume of sessions by channel, engagement, the click-through rate of SEO pages, the cost per lead for SEA and the value per conversion (revenue, average order value, customer lifetime value). The rate itself is read by entry page and by device, and value per visit is compared with the cost of acquisition.

Should the conversion rate be optimised page by page or at site level?

Both, but not for the same use. The site rate serves for tracking and alerting: it says that something is happening, never what. The work itself is done page by page, on a short list of pages chosen at the crossing of qualified traffic and business potential. Below a certain monthly volume, a page rate cannot be read.