SEO vs SEA: allocating budget keyword by keyword and measuring ROI

When a B2B marketing team pays for clicks, the SEO vs SEA debate is settled keyword by keyword: what to keep buying, what to win organically, and how to prove the return of the switch. This guide sets out an allocation table by keyword family, four ROAS metrics to read the real profitability of paid search, a before / after switching method and the place of AI Overviews in the decision, so you can direct paid budget to where organic is hard.

SEO, SEA, SEM and GEO: who does what in search

SEO (organic search) works on natural results: no cost per click, a lead time of several weeks to several months, and rankings that last as long as the page stays relevant. SEA (paid search) buys ads charged per click, mainly via Google Ads (formerly AdWords): visibility is immediate and stops when the budget does. SEM covers both, and in recent usage GEO is added to target the answers of AI engines.

The following table sums up what separates the four terms when you need to make a trade-off:

The split of clicks shows why organic weighs in the decision: 70% of clicks go to organic results (SEO.com, 2026). SEA keeps speed and control of the message; SEO keeps credibility and a cost that does not rise with every additional visit.

SEM: an umbrella term, not a Google Ads budget

In many acquisition teams, "SEM" is used as shorthand for the Google Ads budget. This confusion has a cost: you under-invest in the organic foundation, which reduces dependency on CPC, and you lose unified management of search intent. Read as a whole, SEM is about deciding where to pay to accelerate and where to invest to compound value.

What organic search delivers over time

Organic search builds an asset: pages that keep generating visits and leads without paying for every click. Over time, the marginal cost of a visit falls, and organic keeps a credibility advantage with users who ignore ads. An SEO plan should therefore be judged on overall ROI, whereas SEA is read click by click.

Allocating keyword by keyword: the right unit of decision

Pitting "SEO" against "SEA" as a whole leads to average decisions. The useful unit is the keyword, or the group of queries sharing the same search intent: for each one, you look at the organic position, the cost of the paid click, the presence of competitors in the ads and the value of conversions. Four query families cover most B2B cases:

Cannibalisation and "stolen" keywords

Cannibalisation means paying for a click you would have earned for free: on your brand, or on a keyword where you already rank first organically with no paid competitor. The opposite case is "stolen" keywords: competitors bid on your name and capture part of the demand that is rightfully yours. Defending with SEA can then create incremental conversions, to be weighed against margin and lead value.

Identifying "winnable" keywords

A "winnable" keyword is one whose organic position can be reached at reasonable cost. It becomes a candidate for reducing paid spend when three conditions are met:

• the organic page is already in the top 3, or close to it, and SEA is not creating measurable incremental conversions;
• an editorial plan (pillar page, comparison, use case) and, where needed, authority work can take over part of the budget;
• the time to rank remains compatible with your commercial objectives; otherwise, SEA stays in place on that keyword.

All organic, all paid or hybrid

For each topic, a simple rule sets the starting point:

• 100% organic: stable, educational topics (definitions, methods, high-level comparisons), long cycles, an authority objective;
• 100% paid: seasonal or short-window offers, a need for immediate volume, a quick test of a positioning;
• Hybrid: most B2B cases; SEA learns fast (messages, objections, segments), then those learnings feed SEO pages that can convert over the long term.

The four ROAS metrics that settle the allocation

ROAS (return on ad spend) compares the revenue generated with the cost of the ads. Calculated across the whole account, it mixes clicks that have nothing in common: your brand, which you would have captured for free, and contested generic terms. Splitting it into four subsets shows where paid search really pays off:

Reading the four metrics together

A flattering global ROAS often hides a very high brand ROAS with cannibalisation: those conversions would have happened without an ad, and the ratio is mechanically inflated. Non-brand ROAS is the most honest metric for judging conquest. Set against organic positions keyword by keyword, it shows where to save on SEA and where to reinforce it, because organic remains difficult there.

These four metrics are the ones calculated by the Incremys SEO vs SEA module, presented at the end of this article, from connected Google Ads, Search Console and Google Analytics 4 data.

Budget: balancing short term and long term

There is no universal ratio between SEO and SEA. Three parameters set the split: the marginal cost of the paid click in your sector, the SEO maturity of the site and the expected speed of results, over a quarter or a year. The more immediate the need, the more SEA weighs; the more mature the site, the more SEO can take back control.

A paid portfolio in four buckets

On the SEA side, the most robust split takes the form of a portfolio:

• Tests: a budget reserved for learning (messages, segments, landing pages); without it, you optimise blind;
• Non-brand conquest: capturing explicit demand on needs and solutions;
• Brand and protection: to be decided case by case, between incremental conversions and cannibalisation;
• Retargeting: recovering intent that is already engaged, with a frequency cap and lead quality control.

This portfolio also produces data that is useful for SEO. A topic with a high CPC and strong conversion becomes an SEO priority: every organic position gained there reduces dependency on cost per click. A highly seasonal topic, by contrast, justifies speeding up editorial production before the peak.

Criteria for stopping, restarting and reallocating

To stay in control, three criteria written in advance prevent ad hoc decisions:

• Stop: SEA brings no measurable incremental conversions on a segment already dominated organically, notably the brand;
• Restart: an algorithm update or a change in the results page (AI Overviews, new formats) makes the organic click share drop;
• Reallocate: queries become winnable in SEO while their CPC rises; budget shifts progressively towards keywords that are difficult organically.

Profitability benchmarks, without a like-for-like comparison

Google Ads campaigns show an average ROI of 3.7 (Start'Her, 2026). This benchmark helps you position your campaigns and must be recalibrated for your sector and sales cycle; CPC, CPA and conversion rate benchmarks are gathered in our SEA statistics.

On the organic side, SEO shows an average ROI of 5.1 on long-term objectives (Odiens, 2025). The two benchmarks cannot be compared directly: the first applies to fast-conversion objectives, the second to several years. The method for estimating the return of an editorial plan is detailed in our guide to SEO ROI.

The progressive switch: reducing SEA one query group at a time

Cutting paid spend on a single keyword, all at once, proves nothing: volumes are too low and seasonality blurs the reading. Media buying does not directly influence organic rankings either (Google Search Central): cutting a campaign does not lose positions, and keeping it does not lift the page. The switch is therefore run by groups of queries with the same intent, in six steps:

• Choose the group: homogeneous queries, a single intent and one reference SEO page;
• Check the signals: organic page in or near the top 3, stable traffic, observed conversions, low measured contribution from paid on these queries;
• Measure the before: paid cost, clicks and conversions, organic impressions, CTR and average position, over a reference period;
• Reduce progressively: lower the group's budget in steps, over a comparable period, rather than cutting it outright;
• Compare the after: total conversions, total cost and organic share of the group's demand;
• Decide and log: continue the reduction, stabilise it or restart, and record the decision with its figures.

If total conversions hold while overall cost falls, the switch is rational. If they drop, paid search was incremental on that group, or the organic position was not yet stable enough: you restart, then try again later.

When to keep buying

The switch is not the only outcome. Three situations lead to different choices:

• Keep buying when the SEO time to rank is incompatible with the commercial objective, or when the results page is very unstable;
• Switch progressively when organic holds the top 3 and the contribution from paid is low;
• Keep a safety net during high-stakes periods (seasonality, launch), then reassess once the situation has stabilised.

Measuring: what each data source covers

No single source sees both organic and paid. Cross-referencing the following three is the condition for allocating on facts:

The CTRs of the two channels can only be compared on an equal scope: same query, same device, same period. Nor does a correlation between lower paid spend and higher organic traffic prove causation; only a test isolates the effect.

Incrementality and last-click bias

Incrementality refers to the additional conversions obtained thanks to paid search. It is measured with a test with and without delivery: you cut or reduce ads on a group of queries and compare total conversions, not just those attributed to SEA. It is the only way to tell a budget that creates demand from one that merely captures it.

The last-click model distorts the reading in B2B. The final conversion often goes through a branded search, captured by an ad, even though several organic pieces of content prepared the decision. Conversely, an SEO page can appear high performing because it absorbs demand created elsewhere. Combining assisted conversions with long analysis windows, aligned with the sales cycle, gives a consistent reading, if not a perfect one.

AI Overviews and AI Mode: a new allocation parameter

At the time of writing, AI Overviews and AI Mode have been available in France since summer 2026. These answers generated at the top of the page can satisfy the user before any click, which lowers organic CTR at an equal position. They also change the paid-search equation: 95% of queries with an AI Overview show no ads (Semrush, 2025).

On these queries, the most visible spot cannot be bought: you need to be cited as a source, which is the domain of GEO (Generative Engine Optimization). Google is also testing ads in AI Mode, but the format is not stable at the time of writing. The presence of an AI Overview therefore becomes an allocation input, just like position and CPC.

Tracking AI presence keyword by keyword

For each high-stakes keyword, three questions complete the allocation table:

• does an AI Overview appear, and is your site cited in it as a source?
• is organic CTR falling at a constant position, a sign that the AI answer is capturing part of the demand?
• are organic conversions holding despite fewer clicks, because the remaining visitors are more qualified?

If clicks and conversions fall together on a high-value keyword, the restart criterion comes into play, after first checking that ads still appear on that query. In parallel, the content is reworked to be cited.

Launch, new site, international: adapting the mix

A few situations come up in most B2B allocation decisions:

• Offer launch: SEA generates signals quickly (clicks, leads, objections), while SEO builds the content foundation that makes acquisition sustainable;
• New site: SEA avoids waiting for indexation and early rankings, while SEO builds the structure, internal linking and content plan;
• Competitive market: SEA targets niches and differentiating messages, while SEO works the long tail and distinctive angles (case studies, data, expertise).

Internationally, the logic stays the same. SEA speeds up entry into a new market, but building local assets (pages in the market's language, adapted proof points, internal linking) is what then stabilises acquisition cost.

Synergies, click-through rate, how paid search works: the related guides

The relationship between SEO and SEA details the learning loop: how campaign search terms feed the editorial plan, how to build hybrid landing pages and how to organise a test / conquer / defend portfolio.

The article on SEA vs SEO click-through rate explains how to compare the two CTRs on an equal scope, read the CTR curve by position and diagnose a page that combines high impressions with a low click-through rate.

The guide to Google paid search describes how the auction and Ad Rank work, what a keyword really costs and how to measure a campaign's incrementality before replacing it with SEO.

Depending on your situation

Your starting point sets the first action:

• You pay for your brand with no competitor: isolate brand ROAS with cannibalisation and test a reduction on that group.
• Your generic CPC is rising: identify winnable keywords and launch a progressive switch on a first group.
• Your site is new: keep SEA on decision queries and fund the SEO foundation now.
• Your organic clicks are falling at an equal position: check for AI Overviews before restarting paid search.
• You also advertise on social networks: the Meta advertising training covers media buying outside search engines.

To compare organic and paid performance keyword by keyword, the SEO vs SEA module of the Incremys GEO & SEO platform combines Google Ads, Search Console and Google Analytics 4, calculates the four ROAS metrics and the advertising cost equivalent of your organic positions, to direct paid budget towards keywords that are difficult organically.

FAQ

What is the difference between SEO, SEA and SEM?

SEO works on natural results, with no cost per click but with a lead time, and produces lasting visibility. SEA buys pay-per-click ads, visible immediately and stopped when the budget stops. SEM covers both; in recent usage, GEO, which targets the answers of AI engines, is added to it.

SEO or SEA for an SME?

Both, in different roles. An SME on a tight budget keeps SEA on a few profitable decision queries and invests the rest in an SEO foundation on its winnable keywords. Paid search brings leads straight away, along with learnings; organic search then reduces dependency on cost per click.

Should you bid on your own brand?

Only if competitors are bidding on it. When you rank first organically and are alone on your name, those clicks are cannibalisation: you pay for visits you would have had for free. When competitors buy your brand, a defence can protect conversions; a test with and without delivery confirms it.

How should you split budget between SEO and SEA?

No universal ratio holds. The split depends on the cost per click in your sector, the SEO maturity of the site and the time horizon for results. A robust method funds SEA on immediate demand and keywords that are difficult organically, an SEO foundation on winnable keywords, then rebalances group by group.

Can you stop SEA once SEO is secured?

Yes, progressively and group by group. When the organic page holds the top of the page and paid search brings no incremental conversions, you reduce the budget in steps while tracking total conversions. You keep a safety net during high-stakes periods and restart if the results page changes or clicks drop.

How do you prove a switch has worked?

By comparing before and after on a homogeneous group of queries, over a comparable period. If total conversions hold and overall cost falls, the switch has succeeded. If conversions drop, paid search was incremental. Looking only at conversions attributed to SEO or SEA distorts the conclusion.

Do AI Overviews change the allocation?

Yes, on the queries where they appear. They reduce organic clicks at an equal position, and most of them show no ads: the most visible spot cannot be bought there. You then need to aim to be cited as a source and track, keyword by keyword, the presence of an AI Overview before deciding.

How long does it take to see results from SEO?

Several weeks to several months, depending on competition, site authority and the resources invested in content, technical work and authority building. SEA, by contrast, drives traffic as soon as the ads go live. This gap justifies keeping paid search on a keyword until the organic position is secured.