SEO conversion rate: measuring the one your organic traffic produces

You have an SEO conversion rate in front of you, and a single question: does it mean anything? Your tool often displays several versions of it depending on the screen you open, and last month’s variation may come from the traffic, from the pages, or from the measurement itself.

This page is for the manager who has to set up a stable organic measurement (unit, period, definition of the conversion, scope, exclusions), then read its variations without making them say more than they do. If your question is rather about the share of your conversions that belongs to organic, look first at what SEO conversion covers.

What an organic conversion rate measures

Analysing a conversion rate starts with segmenting the acquisition channels, at the very least SEO on one side and SEA on the other. The SEO conversion rate covers the organic traffic segment alone, and that segment has to be defined: it groups the visits arriving from an unpaid search engine result. Excluded from it are:

• paid visits, even when they land on the same pages;
• visits typed straight into the address bar;
• visits coming from a link placed on another site.

Two edge cases are settled by each organisation: brand queries, which feed the segment without any acquisition work having produced them, and visits born of a generative answer, whose origin is often lost along the way. What matters is not the answer chosen: it is that it is written down and does not change again.

The rate used is the one that underpins a good conversion rate: conversions set against sessions, on the “Organic Search” channel alone. It is written (Number of organic sessions that fired at least one key event / Total number of organic sessions) x 100, the key event being an action linked to revenue, such as an order or a qualified contact request.

The numerator counts sessions, not conversions: out of 100 sessions of which 8 lead to a purchase for 10 orders placed, the rate is 8%, not 10%. For example: if the “Organic Search” channel generates 2,000 sessions in a month and 60 of them convert, your SEO conversion rate is (60/2000) x 100 = 3%.

The figure is easy to produce, and that is what makes it misleading. The two terms of the ratio are choices, not observations:

• in the numerator, the actions you decided to count;
• in the denominator, what you decided to call a unit of traffic;
• between the two, a rule which, if it changes, moves the result without a single sale having been won or lost.

Setting the measurement before reading the figure

An organic rate is read every month, in committee, sometimes by people who were not there when it was defined. Produced by a stable, written rule, it compares with itself over time; produced by factory settings, it compares with nothing. The rule fits in one sentence: for a reliable analysis, keep the same definition of “conversion” and the same unit, the session, across the whole period.

Sessions, users, clicks: three denominators, three figures

In the same month, on the same site, three denominators give three different figures. A visitor who comes back three times counts as three sessions and as a single user: dividing by users therefore produces a higher rate, on identical conversions. Search Console, for its part, counts clicks before arrival on the site. The table compares what each unit counts, what it leaves out and the use it is suited to.

The session is the obvious choice: it is the unit measurement tools produce by default and the one whose history most often remains available as it is. On a long sales cycle, a count by users can inform the reading if it is presented separately and named: being structurally higher, it otherwise looks like a performance.

What gets set once: the definition of the conversion, the period, the exclusions

Before the first figure, five decisions are written down and do not move again, on half a page of internal documentation:

The definition of the conversion, first. If a quote request and a white paper download share the same counter, the rate rises as soon as your content works and no longer says anything about your sales. Write the list down, keep it short, and track micro-conversions separately: they never enter the rate.

The period, next. The reference history covers at least 12 months, one seasonal cycle, read on a rolling monthly basis. The reading cadence is set with the other conversion rate KPIs; the window on which you decide is set by the volume of conversions, not by the calendar. On a low volume, gaps are nothing but sample noise:

• five conversions instead of eight on three hundred sessions make the rate lose a point without anything having happened;
• shortening the window does not make the measurement more responsive, it makes it noisier;
• on a long sales cycle, how long a visit stays credited with a later conversion is set on the real cycle, never on the rhythm of the reporting.

The exclusions, last. Four categories are removed, and their list is written in the same place as the rest:

• internal traffic;
• identified bots;
• test conversions;
• obviously off-target requests.

An undocumented exclusion is the first thing forgotten, and the first thing that explains a gap six months later.

The scope is the worst-framed decision, because it looks as though the tool takes it. The channel group that is authoritative is the one in your measurement tool, and that one alone. Three slippages blur it:

• Search Console provides no channel group and stops at the click: it is read alongside the rate, never in the same column;
• an organic link opened from a messaging app frequently arrives as referral or direct;
• a visit born of a generative answer often files itself somewhere other than organic.

Your scope is therefore a convention: name it, and hold it.

Reading a variation: traffic, pages, or measurement?

Your organic rate has moved by 0.8 points. Before looking for a cause, ask whether it is the traffic, the pages or the measurement that has changed. The flattering explanation is only a hypothesis, checked by reopening both terms of the ratio. The three families of causes do not call for the same actions:

When it is the traffic that has changed

This is the most frequent family, and the least worrying. Lasting organic traffic remains subject to fluctuations, tied in particular to:

• seasonality;
• competition on your queries;
• content freshness;
• all the factors that weigh on SEO.

A variation of that nature is not corrected, it is observed. It still must not be labelled seasonal by default: the hypothesis is checked by finding the same dip on the same date in previous years.

Two causes move the rate without touching a single page. The brand / non-brand mix first: queries carrying your name convert at a level unrelated to the rest, and a week of brand awareness pushes up the organic rate of the whole site. The arrival of new entry pages next, which changes the composition of the traffic before it changes its quality. The right question then becomes: who came in this month who was not coming in last month?

When it is the pages that have changed

The pages family becomes visible as soon as you stop looking at the aggregate figure, because an overall rate mixes pages that are not expected to do the same thing:

• a page newly ranking on a less commercial query attracts a volume that converts less: the overall rate falls while nothing has got worse;
• losing positions on your informational content, more awkwardly, pushes the rate up;
• more broadly, any change in the share of decision pages in the traffic moves the overall rate.

Reading the rate by entry page, or at least separating decision pages from the rest, usually settles it.

When it is the measurement that has changed

This is the family looked at last, when it explains the sharpest breaks: a rate that drops all at once nearly always comes from the counter. Four causes cover the essentials:

• a change of tool or of tracking version, which redefines what a session is along the way;
• an unintended change to the definition of the conversion, when a new form fires the same event as an old one;
• a change in consent collection, which removes visits from the denominator without removing the matching conversions;
• a filter added and then forgotten.

Hence a non-regression check, to be run before announcing anything at all, with every change of tool, of tagging or of consent banner:

• run both configurations in parallel over a full month where that is possible, to get two values actually measured over the same period;
• otherwise, do not recalculate the history: open a new series at the date of the change, keep both side by side and flag the break on the chart;
• never subtract a gap assumed to be constant, because the correction would manufacture a history that nobody recorded;
• keep a change log alongside the dashboard, one line per change with its date: it is the only document that makes a history interpretable.

When the reading stops and the action begins (which pages to work on, in what order, on what volume of conversions to decide), what follows belongs to the method for optimizing the conversion rate.

What generative AI changes in the denominator

The denominator of an organic rate is contracting, and it is not visible in the rate. Generative answers answer in place of the result, and the click that fed your measurement no longer happens, even at the top of the results. The click-through rate itself, position by position, is the subject of optimising CTR in the SERP.

On the volume side, the fall in organic traffic linked to the arrival of generative AI is estimated in a range of -15% to -35% (SEO.com, 2026; Squid Impact, 2025). The range is part of the measurement: it covers very different situations depending on the queries, detailed in the GEO statistics.

The effect on your measurement is arithmetical, and counter-intuitive:

• the denominator loses the least engaged visits first, the ones that came looking for information and left with it;
• the numerator, for its part, holds longer;
• your rate therefore rises while your conversions stagnate, or even fall back.

The movement is not unfavourable in every respect: users who consult an AI answer before clicking arrive with a more precise intent, and that part of the rise is not attributable to your work.

An organic rate rising on a falling volume is not announced as an improvement until the conversions have been looked at in absolute value. They may have held, if the traffic lost was the traffic that was not converting, and that case is stated too. The rate is therefore never commented on alone; it is always displayed with:

• the number of organic sessions;
• the number of organic conversions, in absolute value;
• a mention of any break in the series over the period.

That leaves the search that ends without a click: your brand can be cited, read and remembered there without a single line appearing in your tool. Your organic rate now measures whatever remains of your search traffic, and no setting will close the gap.

What an organic rate does not prove

A proportion is not a progression. An organic rate going from 2.4% to 2.9% does not prove that your SEO is improving: it proves that the ratio between two numbers has changed, for reasons that do not necessarily have anything to do with your work.

Two rates only compare if they were produced by the same rule. So the following do not compare:

• a rate measured on sessions and a rate measured on users;
• your rate this year and last year’s, if the definition of the conversion has changed in the meantime;
• your rate and a published average whose numerator, population and exclusions you do not know.

To place your figure anyway, the order is fixed: your own history at a constant definition first, your segments next, and last the published benchmarks with an identical denominator, as an order of magnitude and never as a target.

Nor does the percentage contain the value of a conversion: two sites on the same rate do not earn the same thing, and it is that value, set against the acquisition cost, that settles a budget trade-off. Finally, the checkout funnel, usability, speed and tagging cannot be fixed from the content side or from the positions: those workstreams have their own owners.

Continue reading

Depending on what you need to do with your measurement, four reads extend this one:

• Your rules are set and have to hold inside the tool. Declaring conversion events, filtering the right channel group and checking the reports: that is the subject of the conversion rate in Google Analytics.
• Your measurement is stable and you are looking for an outside benchmark. It is taken where it is recorded, named and dated, with its denominator: see the conversion rate by business sector.
• You sell B2B. Published benchmarks there do not always count the same thing as your rate: see the average B2B conversion rate.
• A page is dropping off with no traffic or measurement cause. The fix belongs to a discipline with its own methods and vocabulary: conversion rate optimisation.

The performance reporting module crosses organic ranking data, AI citation data and conversion data: it is that crossing which makes a rate readable when its denominator is moving.

Frequently asked questions

What is the SEO conversion rate?

The SEO conversion rate measures the share of sessions from organic traffic that lead to an action linked to revenue, such as a purchase or a quote request. It is calculated separately from paid traffic. Its value depends on two choices: the actions counted in the numerator and the unit of traffic in the denominator.

How do you calculate an SEO conversion rate?

The SEO conversion rate is calculated by dividing the number of organic sessions that fired at least one key event by the total number of organic sessions, then multiplying by 100. The difficulty lies in holding the same conventions from one month to the next: the same definition of the conversion, the same unit.

Over what period is an organic conversion rate read?

An organic conversion rate is read over a history of at least 12 months, to cover one seasonal cycle, on a rolling monthly basis. The decision window is set by the volume of conversions, not by the calendar: on a low volume, a few conversions fewer move the rate without anything having happened.

Why is my organic rate rising while my conversions are falling?

The organic rate rises when the denominator contracts faster than the numerator. Generative answers absorb informational searches first, and therefore the visits that converted little: the ratio climbs on a falling volume. A rate rising on falling traffic proves nothing on its own; always present it with sessions and conversions in absolute value.

What is a good SEO conversion rate in B2B?

A good SEO conversion rate in B2B is one that makes the acquisition cost compatible with the value of a lead multiplied by your closing rate, and that progresses against your own history. Published B2B benchmarks by sector place your figure without setting a target, provided you read their denominator. Then connect organic conversions to the pipeline (MQL, SQL).

Can GEO and LLMs improve the SEO conversion rate?

Yes, indirectly: ranking well in SEO raises the chances of being cited in AI answers, which strengthens trust and brings visitors in with a clearer intent. Content that provides proof mainly serves decision pages. Read that rise together with the volume, though: when traffic contracts, the rate goes up without a single extra conversion.