Is investing in SEO? The 3 numbers that prove it

A manager at an education institution came back from an industry event with one urgent idea: the brand had to show up in AI searches.

Right there, the team typed a generic question about one of the programmes into Google. The first and third links cited in the AI answer were posts from the institution's own blog.

That scene sums up the whole debate about whether investing in SEO is worth it. The novelty the manager brought back from the event was already delivered, already paid for and already working in-house. The problem was never that SEO did not perform, it was that nobody with decision power was looking at it.

The story repeats in companies across every sector. Agency owners conclude SEO does not sell, managers cut organic to boost paid media, and the channel that sustains long-term recruitment slowly disappears from the budget without anyone running the numbers.

You do not need 10 reasons to invest in SEO, you need three calculations, and this article runs each one with figures and sources. In the end, the decision to invest or cut is still yours, but it stops being a guess.

Who declared SEO dead?

SEO did not die for lack of results, it left the stage for lack of an advocate. The drop in clicks is real and documented, but the death-of-the-channel thesis is born in two specific places: at the agency that stopped offering the service, and in the management that stopped looking at it.

Caption: investing in SEO stopped being a contest for position and became a contest for citation inside the generated answer

The first profile is the agency owner who sells SEO and paid media in the same portfolio but only defends one of them in the sales meeting. After months of not offering it, they conclude SEO has no demand. A product hidden at the back of the shelf never had demand, in any market.

The second profile lives inside companies and education institutions: the CEO and the marketing team responsible for student recruitment. Target pressure pushes budget toward the channel that delivers tomorrow, and the channel that delivers for years becomes the line to cut.

And that is precisely the cut that should not happen. Organic and paid have to move together, because it is the sum of the two that brings cost per enrollment down: paid sustains today's volume, and organic makes tomorrow's recruitment cheaper.

There is a structural aggravating factor: the SEO specialist sits in production, and somebody else presents the project to the client. When the person selling does not understand what they are selling, the service vanishes from the proposal, and the result that keeps happening is left without a spokesperson.

What unites both profiles is the absence of grounding. The decision to abandon SEO rarely comes with a cost comparison between channels, attribution measurement or a long-term projection. The cut is born of impression, and the public numbers point the other way.

The true part of the thesis deserves to be stated with a number. A Pew Research Center study of 900 US adults, from searches made in March 2025, found a click on a traditional result in 8% of visits with an AI summary, against 15% of visits without one. Clicks did shrink.

But fewer clicks does not mean a dead channel, it means a new shop window. Anyone concluding otherwise stopped the analysis at the first metric, and we have already shown with data what AI changed in search, separating what ended from what moved.

Paid media is rent and SEO is an asset: what is the difference in CAC?

The difference between paid media and SEO is the difference between renting and building. Investing in media buys presence while the invoice is paid; investing in SEO builds an asset that keeps generating demand years later. CAC behaves in opposite ways across the two channels.

Paid traffic works as an auction: the advertiser competes for every click, and the competition only gets more expensive. WordStream's benchmark report, built on 17,998 search campaigns in the United States between April 2023 and March 2024, recorded a cost-per-click rise in 86% of the industries analysed, with cost per lead up around 25% over the period.

This is where SEO becomes, over time, the cheapest channel in the house. The investment happens once and the return repeats. The article published today serves this month's lead, next year's and the one three years out, dividing the original cost across every new visit. It is the closest thing to compound interest digital marketing offers.

In numbers, the squeeze of the auction looks like this:

Table: Google search ad benchmarks that weigh on the decision about investing in SEO (WordStream/LocaliQ, US campaigns, April 2023 to March 2024; figures are medians, in dollars)

The education line deserves attention from anyone recruiting students. With the sector's cost per lead above the overall average, every CPL increase arrives multiplied at cost per enrollment, because between the lead and the enrollment there is still a whole funnel of losses.

And the logic is not exclusive to education. Anyone researching why to invest in SEO for e-commerce or services finds the same auction getting more expensive: the name of the conversion changes, not the behaviour of the cost.

Organic moves the other way. BrightEdge's reference research, from 2019, put organic search at 53% of traffic across the sites analysed, and that traffic does not charge per click. The figure predates AI-generated answers, and it serves precisely to show that the case for organic never rested on click volume: it rests on cost per visit.

Here is how the two behaviours line up side by side:

Table: cost and permanence compared between paid media and investing in SEO

The table does not condemn paid media, which solves the quarter's target like no other channel. The problem is the full swap: an operation living only on the auction stays hostage to a cost that rises every year and accumulates no digital patrimony at all.

A recent diagnosis we ran at a higher education institution captured that swap. The site published a paid campaign landing page almost every day, while the blog, with no active content planning, had been idle for over a year. All the energy in the rent, none in the asset.

So the first calculation before cutting is simple: how much of your recruitment would survive three months with no media budget? If the answer is almost none, the problem is not that SEO is expensive, it is that the operation has no reserve.

Video: the conversation that comes before deciding about investing in SEO: paid media and organic content competing for the same budget, on the mkt4edu channel (in Portuguese)

Who gets credit for the conversion SEO started?

The second calculation is attribution, and it explains why SEO looks like it does not convert. In last-interaction reports, the channel that gets the credit is the last click before the conversion, almost always an ad or a branded search. The path that started in organic disappears from the number.

The lead's real journey rarely fits in one click. The person discovers the brand in an article, comes back through a remarketing ad, searches the company name and converts. In the report, paid media closes a sale on its own that content opened.

There is a second kind of waste, less visible: the organic lead that arrives and dies sitting in the CRM (Customer Relationship Management). With no active email flow, no A/B test, no nurturing cadence, the lead goes cold, and the blame falls on the channel that generated it.

Does SEO convert?

SEO on its own does not convert, because its job ends when the right person reaches the door. What turns that visit into a sale is the rest of the operation: the nurturing flow, the sales team, the page that receives the click. Plenty of people hire SEO expecting enrollments, and that expectation is wrong from the start.

If the lead does not become a sale, the right question is not what SEO got wrong, it is what happened after the click. Conversion is the work of a complete digital marketing strategy: SEO opens the path and the other teams close. Expecting one channel to do both ends is building half a funnel.

That also changes how you measure each channel's ROI (Return on Investment). Measuring ROI in paid traffic is easy, because cost and conversion appear in the same dashboard.

Measuring SEO ROI means adding what organic generates to what it saves: every click that arrives at no cost is a click paid media did not have to buy. That saving rarely enters the report, and it is what makes the channel look more expensive than it is. The analysis of when SEO is worth it details that calculation.

The opening scene shows the same phenomenon from the other side. The organic result existed, it even appeared inside the AI answer, but it did not exist in the reports the decision-maker followed. A result with no record becomes an invisible result, and budget does not go to the invisible.

The second calculation, then, comes down to one question: when somebody sees your ad, searches your brand and converts, which channel takes the credit in your report? Anyone who cannot answer that has no basis yet for declaring SEO does not perform.

Why does AI search only cite those who invest in technical SEO?

The third calculation is the most current: visibility in AI answers starts with technical SEO. Google's documentation on AI features in Search is direct in saying there are no additional technical requirements, no AI text file and no dedicated schema, and that a page has to be indexed and eligible for a snippet to be cited. Whoever abandoned the base went invisible in the new shop window.

The same documentation adds a caveat that heads off easy promises: AI Mode and AI Overviews may use different models and techniques, so the set of cited links varies between the features. There is no single position to win.

There is also no paid media that buys a citation inside an AI assistant's answer today. The space is taken by built authority: consistent content, clean technical structure, structured data and upkeep. It is classic SEO, from technical and content work to local SEO, operating in a new layer.

And that new traffic tends to be worth more, not less. Semrush estimates visitors from AI search converting 4.4 times better than the traditional organic search visitor, because they arrive with the question already qualified by the answer. It is the company's own estimate, with no published methodology, so treat it as an indication.

The diagnosis of that institution with the idle blog shows the scale of typical neglect. Technical health at 54%, 94% of pages with no meta description and no structured data for programmes, with a catalogue of almost 200 programmes to promote.

The most revealing detail: no AI crawler was blocked on the site. ChatGPT, Gemini and the rest could come in freely, and the open door made no difference, because there was nothing in order inside to extract.

And a site's health is not a photograph, it is a film. A live site changes every day: a new landing page, a programme taken down, an improvised redirect, a template swap. Every change made without looking at SEO opens a silent error, and with no continuous monitoring, health declines unnoticed until traffic sends the bill.

Trying to cut corners makes it worse. A recruitment site generated in a rush by AI, with no SEO work behind it, delivers exactly the kind of page search engines learned to discard.

Google's policies against content produced at scale to manipulate rankings apply equally to automation, to humans and to a combination of the two. The company projected a 40% reduction in low-quality content and, at the end of the rollout in April 2024, reported that users were seeing 45% less low-quality, unoriginal content in results.

Saving on SEO to publish cheap volume costs twice: the page does not rank in traditional search and offers nothing trustworthy for AI to cite. Performance drops, budget rushes to paid media to compensate, and cost per enrollment climbs another step.

Automation does not replace that work either, and we have explained why SEO in the age of AI still requires specialists. The tool speeds up auditing and production, but it does not decide site architecture, does not validate institutional facts and does not build brand authority.

The third calculation closes the reasoning of the other two. The place where the 2026 customer starts researching is fed by exactly the channel that left the budget, and every month of neglect is a competitor taking the citation that could have been yours.

Does an AI-built site with no SEO make recruitment cheaper or dearer?

An AI-built site with no SEO makes recruitment dearer rather than cheaper. The page is ready in days and looks like a saving, but it is born with no technical base, does not rank, is not cited by assistants, and all demand generation goes back to the media auction. The saving on the build becomes a permanent acquisition expense.

The move is common in educational marketing teams and repeats in other sectors: the new site comes out of the AI generator looking good and fast, with no meta description, no structured data and no search architecture. Nobody budgeted the foundation, only the façade.

The numbers show the price. An Ahrefs study of 1 million pages, in June 2026, found that pages with under 50% AI content account for 82.2% of top-3 positions, and that content with low or moderate AI levels received 2 to 3 times the impressions of high or very high AI levels. Ahrefs itself treats this as correlation and notes that site authority may explain part of the gap.

The right reading of the study is the one in its own title: the search engine does not punish the tool, it punishes the thin page the tool delivers unreviewed.

There is also a technical detail almost nobody budgets. Many AI site builders deliver pages that depend on JavaScript to display content, and a Vercel analysis from December 2024 showed that none of the major AI crawlers render JavaScript. GPTBot and ClaudeBot do fetch the files, but they do not execute them.

It is worth separating what this hits. Googlebot does render JavaScript, so Google's AI surfaces can still read the page; the ones left out are ChatGPT, Claude and Perplexity. The site sold as ready for the age of AI is born unreadable precisely to the assistants growing fastest outside Google.

The outcome is what this article's calculations already showed: with no organic traffic and no AI citation, the whole of recruitment leans on paid media, the very channel whose cost per lead rose in 19 of 23 industries. The cheap build becomes the expensive acquisition, enrollment by enrollment.

The problem is not using AI to build, it is skipping SEO along the way. A site generator with technical grounding, reviewed content and structured data works. Without that, the institution trades an investment that compounds for a shortcut that charges interest.

Where do you start investing in SEO without abandoning paid media?

The path is not swapping paid media for SEO, it is making both move together. Good paid media management holds the quarter's target while organic builds the asset, and it is that sum, not the swap, that reduces cost per enrollment and cost per sale sustainably.

Four steps organize the start, in this order:

1. An honest site health diagnosis: technical debt and content gaps define the order of work, not a fixed package. Technical errors, accumulated redirects, pages with no meta description and missing structured data form the kind of debt no new content makes up for.
2. Measuring attribution properly: as long as the report gives full credit to the last click, every budget decision will favour the wrong channel, and organic will keep financing conversions it does not sign.
3. Preparing content for the AI layer: a direct answer at the top of pages, identified authorship, standardized institutional facts and structured data implemented.
4. Putting SEO back on the sales table: the internal step, and the cheapest. A service that is not offered does not sell, and a result that is not presented does not defend budget.

Video: six practices for starting out when investing in SEO, led by whatever the site's technical debt demands first, on the mkt4edu channel (in Portuguese)

The first step is also what answers how to invest in SEO in your specific case. It is where serious SEO work starts, before any new editorial plan.

The third step has a scope of its own. It is SEO for LLM, the work aimed at large language models, which decides which brands appear in assistants' answers.

Whoever produces the organic work needs a spokesperson in the room where the decision happens. Without that, the first three steps work and nobody finds out.

Frequently asked questions about investing in SEO

What do you lose by postponing investment in SEO?

Postponing investment in SEO costs more than the budget saved. Every month idle adds acquisition cost rising in the media auction, organic conversions with no record in attribution, and citations in AI answers taken by competitors who kept the basics in order.

This article's three calculations fit into three questions: how much of your recruitment survives with no media budget, who gets credit for the conversion content started, and what AI finds when it visits your site. Anyone who answers all three with data rarely concludes that SEO is dead.

A full SEO health diagnosis answers all three at once. It maps the technical errors holding your positions down, shows where organic visibility is being wasted and measures how your brand appears today in answers from ChatGPT, Gemini and Google's AI features.

With that picture on the table, the budget conversation changes character. Instead of debating whether SEO is worth it, you start debating what to fix first, how much traffic and position that unlocks, and what is missing for your brand to become a citation in AI answers.

That picture exists for your site too, and it fits in a conversation. Talk to our team and request the full SEO health diagnosis, with the three calculations run on your numbers, not on guesswork.