SEO Is No Longer a Performance Channel

Five or six years ago, the SEO industry's rallying cry was "become a publisher." Create content, rank for keywords, watch the traffic roll in. Every agency deck had the same slide: blog more, build topical authority, capture the long tail. And it worked — for a while. I was one of those people. I told clients to publish more, blog more, build topical authority. I believed it. They believed it. The traffic came, the dashboards looked healthy, and we all told ourselves we were doing marketing.

The chickens are coming home to roost. The advice that built traffic for a generation of businesses is now the exact strategy that's being dismantled. Systemically, structurally, and — if you're paying attention to the data — irreversibly.

A few weeks ago I argued that AI search is not a performance channel. The piece landed harder than I expected, which usually means I'd struck something people already suspected but hadn't articulated. Then I showed that only roughly 10% of Google searches represent genuine discovery — the kind where someone types a query without a destination in mind and clicks a result they've never heard of. Most search traffic is people who already know where they're going using Google as a navigation layer. Not a discovery engine, but a directory with delusions of grandeur.

The part that really stings is that we didn't just follow bad advice — we were played. For fifteen years, publishers and brands poured billions into creating content that made Google's search results better. Blog posts, guides, reviews, explainers — all of it indexed, ranked, and monetised by Google through advertising. We gave them the raw material that trained their language models. And now that same content is being regurgitated as AI answers without a penny changing hands. No licensing. No attribution that drives clicks. Just extraction.

This post is the logical extension. If AI search isn't a performance channel, and if traditional search is mostly navigational anyway, then what exactly is organic (Google) search as a growth lever in 2026? The honest answer is: not what you think. If you're betting on organic search to grow your business — not maintain it, not defend existing positions, but actually grow — you're betting on a dead horse. Upper funnel SEO doesn't exist. The search traffic you're optimising for mostly wasn't incremental to begin with. It was demand you already owned being routed through an intermediary that took credit for the journey. And what's left after AI Overviews, zero-click results, and Google's property capture isn't enough to build a business on.

There are exceptions. B2B SaaS and local services still work brilliantly. I'll get to those. But if your typical conversion value isn't in the thousands, if you're a small e-commerce brand betting on search traffic as a revenue growth lever, if you're a publisher, a media company, an affiliate site, or anyone whose model depends on driving traffic from Google to pages monetised through advertising — the ground beneath you has shifted and most of the industry hasn't looked down yet.

I'm going to show you the numbers. Condé Nast planning for zero search. HubSpot losing 70-80% of organic traffic. Chegg abandoning its SEO-first strategy after a 49% collapse. Graphite's analysis of 40,000 websites showing the middle tier getting crushed while the largest platforms gain ground. This isn't a trend. It's a reallocation.

One more thing before we go further. Wayfinder builds technical auditing tools for AI navigation — essentially Screaming Frog for AI. We help businesses understand how AI systems find, interpret, and surface their content. We have planned products that will look more broadly at AI traffic analytics. I believe everything I'm about to write. But I'd be a hypocrite if I didn't tell you where I sit before I tell you what I see. The commercial tension is real: the reallocation I'm describing is, in part, the market we're building for. Read the data and decide for yourself if I'm just a verbose salesman.

---

1. The Hook — Condé Nast Plans for Zero Search

Roger Lynch, CEO of Condé Nast, told his teams last year to plan as if search is zero. Not "plan for reduced search." Not "diversify away from search over time." Zero. This is a man who runs Vogue, The New Yorker, Wired, Bon Appetit — brands with a century of cultural authority — and he's telling his division heads to model a future where Google sends them nothing. ¹

The reason is brutal and simple. At Condé Nast, search had fallen from the majority of visits to roughly 25% over several years. Three consecutive years of search traffic decline, each one exceeding the budgeted drop. When your planning assumptions are wrong three years running, you stop planning for the thing that's declining and start planning for its absence. That's not pessimism. That's arithmetic. ¹

Condé Nast is not alone. The Daily Mail reported an 89% CTR drop from AI Overviews. ² Chegg — the education platform that built its entire content strategy around ranking for academic queries — suffered a 49% decline in non-subscriber traffic and has effectively abandoned its SEO-first approach. ³ Forbes, Forbes, saw 40-50% organic declines despite spending $500,000 per month on an AI-driven SEO operation. ⁴ The pattern is consistent across publishers of every size and specialism.

Condé Nast's strategic response is instructive. Subscriptions grew 29%. Events revenue grew 40%. And crucially, they signed AI licensing deals with OpenAI, Amazon, and Perplexity — extracting value from the same technology that's cannibalising their search traffic. ¹

Notice who Condé Nast didn't sign a deal with. OpenAI, Amazon, Perplexity — all licensed. Google — notably absent. When the platform that built its business on your content won't even pay to license it, that's not a partnership. That's a heist.

I should caveat this heavily. Condé Nast is sui generis. These are 100-year-old brands with subscription-viable content, cultural authority most businesses will never approach, and enough scale to negotiate licensing terms with AI labs. Your travel blog does not have these things. But the direction of travel matters even if the magnitude doesn't. If Condé Nast can't make search work as a growth channel, what chance does a business without Vogue's brand recognition have?

---

2. The Decomposition

To understand why search is failing as a growth channel, you need to look at what search actually is now — not what it was in 2010. The SparkToro/Datos study remains the most rigorous attempt to map this. They analysed 332 million queries across roughly 130,000 US devices from January 2023 through September 2024. The sample is large enough that the directional conclusions are hard to argue with. ⁵

Here's what they found. Approximately 44% of Google searches are branded — someone typing "nike trainers" or "nytimes crossword." Another 33% are navigational — someone typing "facebook login" or "gmail" into the search bar because it's faster than typing the URL. Branded plus navigational accounts for roughly three-quarters of all search behaviour. These people knew where they were going before they opened Google. Google didn't discover anything. It processed a routing request. ⁵

Then there's the zero-click problem. In the US, 58.5% of searches end without any click at all — the answer appears in the SERP, and the user never leaves Google. Of the clicks that do happen, 28.5% go to Google-owned properties. The top 148 keywords alone account for 15% of all search volume. ⁵ SparkToro's latest estimate puts the zero-click rate at 64.82% in 2026, up from roughly 50% in 2019. ⁶

We modelled this quantitatively in our previous post. The Discovery Waterfall — a Monte Carlo simulation of Google's 8.5 billion daily searches — showed that only ~10% represent genuine discovery. The rest is branded search, navigational queries, zero-click results, and clicks to Google-owned properties. The waterfall diagram generated from that simulation is below, but you can also read the full post and explanation here: Google's Biggest Number Is Its Best Lie ⁷

This is not a marketplace of ideas. It's a distribution system with a very small number of winners.

Rigorous estimate: if you subtract branded, navigational, and zero-click from 100%, you're left with roughly 10-15% that could even theoretically be discovery. The exact number depends on definition. I'm not claiming precision here. I'm claiming that the thing SEO was originally built to capture — someone with a problem they don't know how to solve, typing a query, finding your content, becoming a customer — is a shrinking fraction of a shrinking pie.

The "(Not Provided)" Admission

There's another layer to this that most people don't talk about. Google has known for over a decade.

Ten-plus years ago, Google flipped a switch and made all organic search keyword data disappear behind the "(not provided)" curtain in Google Analytics. Their explanation was security — protecting user privacy from third parties. The stated reason was always bollocks. If Google genuinely believed that organic search generated incremental demand that marketers couldn't prove without keyword data, they'd have found a way to provide it. They provide it for paid search. They just don't provide it for organic. ⁸

Here's why that matters. If organic search truly generated incremental demand — if it created customers who wouldn't have found you otherwise — Google would want you to prove it with keyword-level data. The attribution would make the case for Google's importance in the marketing mix. Instead, they hid the data and left a decade-long measurement gap that agencies have filled with inference, modelling, and not-infrequently creative reporting. I know because I've written those reports. Mea culpa.

The fact that "(not provided)" has persisted for 10+ years is tacit admission that Google knows exactly what organic search is and isn't. It isn't demand generation, but demand harvesting. Google takes credit for journeys that would have happened regardless — people typing your brand name, your URL, your product category — and they don't want you to be able to prove that with data. The obfuscation is structural, and it's profitable. For everyone in the SEO measurement chain, the ambiguity is the business model.

---

3. Winner-Take-Most: The Consolidation of Search

The publisher data is grim. But if this were only a media industry problem, I'd be writing a narrower piece. It isn't.

Roger Lynch called it a barbell: large brands and tight niches survive, the middle gets crushed. And from his vantage point — running Vogue, The New Yorker, Wired — that's exactly what it looks like. But to him, "niche" means Pitchfork, the 30 year old website that ranks in the top 7,000 most visited on the entire web. Compared to Vogue? Niche. But from the outside, looking at the millions of websites and businesses on the web, rather than one mammoth publishing house, it doesn't look like a barbell. It looks like a gravity well. The biggest sites pull ever more traffic toward themselves, and everyone else orbits at diminishing distances — or gets pulled into entirely different solar systems.

Graphite and Similarweb analysed 40,000 US websites in January 2026. They found that the top 10 largest sites gained 2% in organic traffic. The smallest sites gained slightly too — Google's helpful content updates and the long-tail's lower AI competition created pockets of opportunity. But the mid-range — sites ranked 100 to 10,000 — declined the most. ⁹

This isn't a barbell. It's a gravity well. The biggest sites pull ever more traffic toward themselves through brand authority, user signals, domain age, and content volume that algorithm shifts can't dislodge. Everyone else orbits at diminishing distances. The smallest sites that appear "flat or slightly up" in aggregate aren't thriving — they're surviving on scraps. The mid-market generalists are being pulled apart.

The B2B data is catastrophic. ZipTie.dev reported in April 2026 that 73% of B2B websites lost organic traffic, with an average decline of 34% year-over-year. ¹⁰ NP Digital studied 50 B2B companies with revenue between $5 million and $100 million and found organic leads down 47%. ¹¹ These aren't publishers. These are software companies, consultancies, industrial manufacturers — businesses that sell to other businesses and have spent years building content libraries to capture "commercial intent" queries. The funnel is leaking at the top.

HubSpot — HubSpot — the company that literally invented inbound marketing as a category, saw organic traffic collapse from 13.5 million to under 7 million monthly visits. ¹² When the cathedral of content marketing loses 70-80% of its search traffic, you have to confront the possibility that the cathedral is falling down. This isn't a failure of execution. HubSpot has a 300-person content team, the best tools money can buy, and decades of domain authority. If they can't hold their ground, the ground itself is shifting.

The local data tells a related story. SOCi's 2026 Local Visibility Index found that ChatGPT recommends only 1.2% of the 350,000+ local businesses it tracked across 2,751 brands. ¹³ 98.8% are effectively invisible to AI search assistants. The reallocation here is from traditional local pack rankings to AI assistants that haven't figured out local intent yet.

The review platform concentration is equally stark. Detailed.com found that 16 media companies rank on page one for 85% of 10,000 product-review keywords. ¹⁴ If you're looking for "best wireless headphones," you're not finding independent bloggers anymore. You're finding Wirecutter, Tom's Guide, TechRadar, Forbes Vetted — the same media brands recycled across every product category. The niche review site that specialised in audiophile headphones can't compete with Wirecutter's domain authority. The game is rigged.

Academic research confirms the structural nature of the problem. Calzada, Duch-Brown, and Gil at the Universitat de Barcelona and CESifo found that Google's core updates systematically reduce news outlet top-10 positions — the algorithm changes don't just shuffle rankings, they shift market share toward platforms and away from publishers. ¹⁵ An NBER working paper from Allcott et al. found that eliminating demand frictions would double Bing's market share, which implies that Google's dominance persists not because it's dramatically better at search but because it's the default. ¹⁶ Users don't choose Google. They don't choose at all.

Where the Niche Creators Actually Went

The smallest sites appear flat or slightly up in some datasets — but look at where niche creators actually are. They're on Substack, Patreon, OnlyFans, TikTok, and Instagram. Not because those platforms are better for discovery, but because Google stopped sending them traffic a decade ago. Niche writers, musicians, adult content creators, indie journalists didn't "survive" in organic search. They were expelled from it and had to build alternative distribution.

These platforms exist because organic search for niche creators is impossible. Substack isn't a luxury for newsletter writers — it's a life raft. OnlyFans isn't a choice for adult creators — it's where they ended up after every other platform demonetised them and Google buried their content. Patreon exists because artists couldn't rank for their own work. The "niche survival" in organic search datasets is largely local businesses and specialist B2B — not the indie publishers and content creators that SEO was supposed to empower.

The Brand Building Thought Experiment

Here's a thought experiment. Name a $100 million+ consumer brand built primarily through organic search in the last decade. Not B2B SaaS — consumer. Not a brand that uses SEO as part of its mix — a brand that scaled to nine figures because of search. I'll wait.

The fact that it's hard to think of one is the point. Gymshark hit $1.3 billion through Instagram and influencers. Glossier hit $1.2 billion through community-led growth. Notion scaled through product virality and community. The brands that grew from nothing without massive media investment did so through social, community, and product — not through ranking for keywords.

Search doesn't build brands. It harvests demand that brands built elsewhere.

---

4. AI Has Eradicated Informational Search as a Growth Tactic

I'll cut off the obvious objection. Sure, it's hard to find a mega-brand built from organic search, but there are brands like Omnisend and Intercom that built massive success from it, and not that long ago, so how can SEO be dead?

It's not. Let me name the specific thing that's dying, because the headline is too broad otherwise.

Informational search — the top-of-funnel, "how do I," "what is," "best way to" queries that the entire "become a publisher" playbook was built to capture — is being systematically absorbed by AI. Not all search. Not transactional search. Not navigational search. Informational search. The thing content marketers spent a decade building libraries around. The thing Omnisend built a $55 million business on. The thing Intercom used to sprint from $1 million to $50 million ARR. That specific thing.

Here's how it worked when it worked. Omnisend, bootstrapped with no venture capital, scaled from zero to 40,000 highly targeted monthly visitors by capturing high-intent terms like "e-commerce email marketing" and "Klaviyo alternatives" — publishing expert-led content that ranked for broad business problems their target customers were already searching for. Content and SEO became their second largest revenue driver globally. Intercom did the same thing from $1M to $50M ARR, publishing multiple times per week on their "Inside Intercom" blog, building what they called a "Semantic SEO" web that captured thousands of informational and transactional queries around customer support, user engagement, product management, SaaS pricing. They didn't wait for users to search for "chat software" — they intercepted them when they searched for foundational business problems like "how to reduce churn" and "product management frameworks," establishing early authority and pushing them down the funnel.

That playbook was the gold standard. It was what every SaaS content marketer was trying to replicate. And it is now the exact playbook that AI Overviews are dismantling.

Google's AI doesn't answer "Klaviyo alternatives" yet — that's transactional, comparison intent, and Google monetises it through Shopping and paid ads. But "how to reduce churn"? "What is product-market fit"? "E-commerce email marketing best practices"? AI Overviews now appear on roughly 35% of informational queries, providing complete answers directly in the SERP. The user who would have clicked through to Intercom's blog post now gets a five-paragraph AI summary that synthesises the best bits from fifty sources. Intercom gets no click. Intercom gets no brand impression. Intercom's content becomes training data for the machine that just replaced their traffic.

This is the specific death. Not all of SEO. Not the search for "plumber near me" or "buy running shoes" or "Intercom pricing" — those survive because Google either monetises them through ads or because the user intent is so commercially specific that an AI Overview can't close the transaction. The death is concentrated in the top of the funnel, in the educational content, the thought leadership, the how-to guides, the industry explainers — the exact content that Google itself now advises you to "create non-commodity content that's helpful, reliable, and people-first" so they can ingest it more effectively.

The distinction matters because it determines what you do next. If you're a B2B SaaS company, ranking for "[your category] software" and "[competitor] alternatives" still works — that's transactional intent that AI can't close without sending users somewhere. But the content strategy of publishing weekly guides to foundational business problems in your industry, hoping to capture top-of-funnel discovery and nurture leads over months? That's the part that's being absorbed. The funnel hasn't collapsed at the bottom. It's evaporated at the top.

Omnisend and Intercom built extraordinary businesses with informational SEO. They are case studies for what was possible. They are also, increasingly, case studies for what is no longer possible. If you tried to replicate either strategy today — publishing comprehensive guides to broad industry problems and expecting organic search to deliver discovery traffic — you would be feeding content into a machine that answers the query before the user ever sees your domain. Not a growth channel. A content donation programme.

---

5. What Organic Search Actually Is Now

If search isn't a growth channel for most businesses, then what is it? I've found it helpful to think in three layers. Call it the A+B+C model.

A: Verification surface. The majority of organic search traffic to brand-name queries is people confirming what they already know. They saw your Instagram ad, heard your podcast interview, got a referral from a colleague — then typed your name into Google to verify you exist. This is valuable. Someone who searches your brand name and clicks through is warm. But it's not incremental acquisition. It's confirmation.

The academic evidence is brutal and old. Blake, Nosko, and Tadelis published their eBay study in Econometrica in 2015 — a decade ago — and found that 99.5% of branded search clicks were retained even when eBay stopped buying ads for its own name. ¹⁷ People searching "eBay" went to eBay. The search engine was a tollbooth on a road users would have travelled regardless. The Brand Science Institute puts it more bluntly: "If people want to find you, they'll find you." ¹⁸ Jim Mollica, former VP Consumer Marketing at Bose: "Are we giving the cash register credit for the sale?" ¹⁹ Viant and the ANA found that 70% of ad budgets go to demand capture, only 30% to demand creation. ²⁰ We've built an industry around harvesting intent we pretend to have generated. Which, when you say it out loud, is mostly bollocks dressed up as strategy.

B: AI citation source. Your content feeds the AI responses that answer user queries, but you don't get the click. Ahrefs found that AI Overview citations from the top 10 organic results dropped from 76% to 38% between early 2025 and March 2026. ²¹ Here's the paradox: being cited in an AI Overview increases clicks by roughly 120% compared to uncited pages. But the absolute traffic is still down dramatically because the AI Overview itself answers the question. A 120% increase on a base that's fallen 80% is not a win.

SE Ranking's analysis of 30,000 commercial keywords in January 2026 makes this concrete. The top five review platforms capture 88% of AI citations — but they've lost 76-92% of their actual traffic. G2 is down 84.5%. Capterra down 89%. TrustRadius down 92.2%. ²² These platforms went from being destinations to being references. I call this the Citation-Traffic Paradox: being cited is better than not being cited, but it's a fundamentally worse commercial position than being the destination.

C: Niche visibility tool. There remain long-tail queries where AI still fails — highly specific technical questions, local intent with unusual modifiers, queries in regulated industries where AI hallucination risk is high. A plumbing company ranking for "emergency boiler repair N1 London" still gets the call. A specialised medical device manufacturer ranking for "ISO 13485 compliant catheter coating specifications" still gets the inquiry. These are real, valuable, and defensible. But they're narrow by definition. You cannot build a venture-scale media business on them.

---

6. Why This Is Happening — Follow the Money

Google's Search revenue was $63.07 billion in Q4 2025, up 17% year-over-year. ²³ Organic traffic to businesses is declining. Paid search revenue to Google is growing. These two facts are not coincidental.

Every time an AI Overview reduces organic clicks, the remaining inventory becomes more valuable — and businesses are pushed toward paid placement. It's an elegant transfer mechanism. Google extracts the same commercial value; it just changes the label from "organic" to "paid."

Nick Fox, SVP and GM of Google Search, said on the February 2025 earnings call: "Google Search is becoming AI search." ²⁴ Google is merging AI Overviews and AI Mode into a single interface where the answer is the product, not the clickthrough. And the commercial logic is taking shape in patent filings. US Patent 12536233B1, filed in January 2026, describes AI-generated landing pages — pages hosted by Google that contain product information, pricing, and direct conversion paths without the user ever leaving the platform. ²⁵

Google doesn't want to be the directory that sends you to the store. It wants to be the store. The search results page was already a tollbooth. The AI landing page is a shopping centre where Google owns the building, collects the rent, and processes the payments. This isn't a conspiracy theory. It's a publicly filed patent, an earnings call statement, and a 17% revenue growth trajectory pointing in the same direction.

Playing the game by Google's rules means playing a game where the house always wins.

And now Google is brazen enough to publish a guide telling you how to play along. In May 2026, Google released "Optimizing Content for AI Search" — an official guide for publishers on how to make their content more discoverable by Google's AI systems. The advice? "Create non-commodity content that's helpful, reliable, and people-first: Focus on developing unique, expert-led content that provides value beyond common knowledge." Read that again. Google is telling publishers to create the exact kind of content that AI can't generate — so Google's AI can learn from it, paraphrase it, and serve it to users without sending them to your site. The content that is most uniquely, authentically yours is precisely the content Google wants most to ingest. It's not a guide to getting traffic. It's a guide to being a better data source for the machine that will replace your traffic. When the platform that built its business on your content publishes a manual on how to give it more content, that's not advice. It's a farm management handbook for the content you're producing free of charge.

This isn't new. The eBay study showed branded search was 99.5% cannibalised a decade ago. ¹⁷ Google has known exactly what organic search is this whole time. They hid the keyword data. They built zero-click features. They launched AI Overviews that paraphrase your content without sending you traffic. And quarter after quarter, their revenue goes up while your organic traffic goes down. That's not a bug in the system. It is the system.

If you're a major publisher or brand, you should consider whether blocking Google entirely makes sense. Make them pay for your content like the AI companies are starting to have to. Use paid ads if you must — at least the value exchange is transparent. You give Google money, they give you placement. That's a market. Organic search hasn't been a market for a long time. It's been an extraction operation, and you're the raw material.

---

7. Counter-Evidence — Where SEO Still Works

I've spent six sections arguing that organic search is in structural decline as a growth channel. Now I need to tell you where I'm wrong. The exceptions are substantial — and they prove the rule.

Reddit's search visibility surged 1,274% year-over-year, driven by Google's preference for UGC in AI Overviews. ²⁶ Quora, Stack Exchange, GitHub — UGC platforms with massive authority have won enormously from the same shift that's crushing publishers. A Reddit thread with 400 comments debating mechanical keyboards under £100 is richer and more trusted than any AI summary.

Programmatic SEO has produced genuine outliers. Flyhomes grew organic traffic by 10,737% through location-specific landing pages. Quillbot added 490% visibility. CK12 drove significant gains by turning educational content into interactive experiences. ²⁷ When programmatic SEO produces a tool or interactive experience that serves a function AI Overviews can't replicate, it works. When it produces 800 words about "best project management software," it doesn't.

B2B SaaS is arguably the strongest remaining case for search investment. First Page Sage's longitudinal study found an average SEO ROI of roughly 702% for B2B SaaS companies, with a 7-month break-even period. ²⁸ A procurement manager searching for "SOC 2 compliance automation platform" is building a vendor shortlist, not looking for a quick answer. The intent is commercial, the content requirements are specific, and conversion values are high enough that ranking for even a handful of terms justifies significant investment. If your typical conversion value is in the thousands, SEO still works brilliantly. If it isn't, the math gets harder.

Local SEO operates in a different environment. BrightLocal found that 76% of searchers who conduct a local query visit a business within 24 hours. ²⁹ A plumber in Manchester doesn't need to worry about Reddit stealing their traffic for "emergency boiler repair M4." The search context is local, urgent, and transactional — exactly the kind of query that still generates clicks.

Here's the critical distinction: if you sell a product or service that people actively search for, with conversion values high enough to justify the investment, SEO still works. If you depend on passive discovery through informational content — listicles, how-to guides, explainers — you're in trouble. Google still sends 345 times more traffic than all AI platforms combined. That traffic is just going to Reddit threads, B2B comparison pages, local business listings, and product pages — not to the 800-word blog post your content team published this morning.

---

8. What Do You Actually Do?

All of this analysis is worthless if it doesn't change behaviour. Here's my best attempt at prescriptive advice.

First, abandon the "replacement channel" mental model. Attention has fragmented across platforms to the point where no single channel replaces organic search. You build a portfolio of surfaces, each capturing a different slice of attention for a different purpose.

VCCP's Hacking the Attention Economy report found that the top 50 advertisers grew revenue by an average of just 3% — while digital ad spend quadrupled from $63 billion to $333 billion over the same period. ³⁰ Ad spend has quadrupled. Effectiveness has flatlined. VCCP estimates up to 66p of every £1 spent on advertising is wasted on impressions that never achieve the minimum threshold for human attention. ³⁰ You are, in many cases, paying for pixels that no brain ever processed.

The generational shift is more dramatic still. Among Gen Z consumers, 46% now prefer social media over search engines for product discovery. Instagram captures 30.4% of product discovery intent among young consumers, TikTok 23.2% — together surpassing Google at 18.8%. ³⁰ If you're planning a content strategy that assumes young consumers start their purchase journey on Google, you're optimising for the behaviour of people who are ageing out of your target market.

A portfolio approach looks different depending on what you sell. Email still returns $36-$45 per dollar spent. Podcasts reach 500 million listeners with an 80% listen-through rate — sustained attention no display ad has achieved. YouTube's 2.5 billion users watch 70% of content on mobile. ³⁰ AI search itself is part of the portfolio. ChatGPT has 200 million weekly active users and 1.6 billion monthly visits. The traffic it sends represents roughly 4.4% of what Google sends — tiny volume, exceptional quality.

The question isn't "what replaces organic search?" It's "what surface captures what kind of attention at what stage of intent?" Brand awareness happens on social and podcast. Product discovery happens on Instagram, TikTok, and YouTube. High-intent commercial queries still happen on Google — for now. Each channel does a different job. The businesses that survive this reallocation stop asking where their search traffic went and start asking where their customers actually are.

By Business Type

If you're a publisher: Look at the Condé Nast playbook. Subscriptions, events, AI licensing, newsletters — surfaces that replace ad-supported search traffic. The 29% subscription growth and 40% events revenue aren't accidents. ¹ Newsletter economics are underrated. A 50,000-person email list at 40% open rates gives you 20,000 reliable impressions per send — more than most mid-tier publishers get from search, and you own the channel.

If you're B2B SaaS: You're an exception. Double down. Invest in topical authority through depth, not breadth. One genuinely useful interactive tool outperforms fifty generic blog posts. If your average contract value is £15,000, you don't need 100,000 monthly visitors. You need 500 of the right ones. ²⁸

If you're a local business: Local SEO plus reviews plus Google Business Profile optimisation is still your highest-ROI activity. But diversify into social presence anyway. Instagram for visual businesses, TikTok for personality-driven services, LinkedIn for B2B local. The 76% visit-within-24-hours stat cuts both ways — local search converts brilliantly, but it's a thin funnel. ²⁹

If you're e-commerce: Social commerce and influencer marketing are non-negotiable. Product pages rank fine. But your blog content is dying. Don't write gift guides. Instagram and TikTok do style content better than you ever will. If you're a small e-commerce brand betting on search traffic as a revenue growth lever, that's foolish. The brands that scaled to £10 million+ without search dependence — look at what they did on social and copy it.

If you're a mid-market brand: Pick a side. Go big — build brand awareness through mass media until you're large enough to survive the consolidation squeeze. Or go niche. The middle is a death zone. The businesses getting crushed in Graphite's data — sites ranked 100 to 10,000 — are mid-market generalists. ⁹ Not big enough for brand gravity. Not small enough to hide in the long tail.

The Agency Incentive Problem

Here's something I don't see discussed enough. SEO agencies still push SEO services because retained SEO is hugely profitable — recurring revenue on retainer hours, month after month, with a client relationship that rarely gets measured against actual revenue outcomes. The connection to ROI has been nebulous for a long time. Feet haven't been held to the fire the way they have in paid search or paid social, where every pound spent can be traced to a click, a conversion, a cost-per-acquisition.

It isn't snake oil — SEO does work. It can generate great ROI. It's just much harder to prove and line up than those channels, and the "(not provided)" gap I described in Section 2 has made that opacity structurally profitable for everyone in the supply chain. Agencies report on rankings, traffic, domain authority — metrics that sound impressive and correlate loosely with business outcomes but can't be tied to revenue the way a paid search CPA can. The measurement gap is the business model.

This is why I think SEO should really be planned more like brand awareness. Accept the measurement limitation and treat it like podcasts, sponsorships, PR, and other broad-stroke content marketing. You know it has value. You just can't precisely attribute which pound generated which pound of revenue. That's fine — as long as you're honest about it. The problem is when businesses treat SEO as a performance channel, demand performance-channel measurement, and the industry pretends it can deliver it. It can't. It never could. The "(not provided)" curtain made sure of that.

Table stakes — technical SEO, on-page, site speed, mobile — spend £20,000 to £50,000 per year and don't overthink it. Strategic SEO — content differentiation, topical authority, conversion optimisation, AI citation engineering — this is where the ROI lives now. Do not invest in: blog content for "organic traffic volume," keyword-stuffed articles, generic how-to guides that AI Overviews answer in two seconds. That money is better spent on almost anything else.

This is not about abandoning search. It's about changing what you optimise for. From traffic volume to commercial intent. From keyword rankings to conversion rates. From "how much can Google send us?" to "how do we build surfaces that don't depend on Google's mood?"

---

We're building an AI Search Absorption Index to track these trends over time — a single metric combining keyword volume distribution, SERP position analysis, and AI overview absorption assessment, built on a pipeline of DataForSEO, SerpAPI, and a locally hosted Qwen model. More on that in the coming weeks.

---

Footnotes

1. Roger Lynch interview — Press Gazette, Digiday, Semafor, 2024. Condé Nast search traffic decline from majority to ~25%; AI licensing deals with OpenAI, Amazon, Perplexity; 29% subscription growth; 40% events revenue growth. ↩ ↩² ↩³ ↩⁴

2. Daily Mail AI Overview CTR drop (89%). Press Gazette, 2024. ↩

3. Chegg non-subscriber traffic decline (49%) and SEO-first strategy abandonment. See: investor.chegg.com ↩

4. Forbes organic declines (40-50%) despite $500K/month AI SEO operation. Search Engine Land, 2025. ↩

5. SparkToro/Datos clickstream study — 332 million queries, 130,000 US devices, January 2023–September 2024. Branded 44%, navigational 33%, zero-click 58.5%, Google properties 28.5% of remaining clicks. See: sparktoro.com/blog ↩ ↩² ↩³

6. SparkToro 2026 zero-click update — 64.82% of Google searches end without a click, up from ~50% in 2019. See: sparktoro.com/blog ↩

7. Wayfinder — "Google's Biggest Number Is Its Best Lie: The Discovery Waterfall." Monte Carlo simulation of 8.5 billion daily Google searches. See: wayfinderai.tools/blog/googles-biggest-number-is-its-best-lie ↩

8. Danny Sullivan, "Google To Begin Encrypting Searches & Outbound Clicks By Default With SSL Search," Search Engine Land, October 18, 2011. See: https://searchengineland.com/google-secure-search-not-provided-119907 ↩

9. Graphite/Similarweb analysis — 40,000 US websites, January 2026. Top 10 sites +2%, smallest flat/slight gain, mid-range (100-10,000) largest declines. Graphite & Similarweb analysis, January 2026. ↩ ↩²

10. ZipTie.dev B2B organic traffic study, April 2026 — 73% of B2B websites lost traffic, 34% average decline. See: ziptie.dev ↩

11. NP Digital B2B study — 50 companies ($5M-$100M revenue), 47% organic lead decline. See: npdigital.com ↩

12. HubSpot organic traffic collapse — 13.5M to under 7M monthly visits. Search Engine Land, 2025. ↩

13. SOCi 2026 Local Visibility Index — ChatGPT recommends 1.2% of 350,000+ local businesses across 2,751 brands. See: soci.io ↩

14. Detailed.com — 16 media companies on page one for 85% of 10,000 product-review keywords. See: detailed.com ↩

15. Calzada, J., Duch-Brown, N., and Gil, R. — "Google's Core Updates and the News Industry: A Longitudinal Analysis." Universitat de Barcelona / CESifo. See: https://papers.ssrn.com/sol3/papers.cfm?abstract_id=4527684 ↩

16. Allcott, H., et al. — "Weak Market Definition and the Behavioral Origins of Corporate Market Power." NBER Working Paper. ↩

17. Blake, T., Nosko, C., and Tadelis, S. (2015) — "Consumer Heterogeneity and Paid Search Effectiveness: A Large-Scale Field Experiment." Econometrica 83(1), 155-174. eBay branded search retention: 99.5%. See: https://www.nber.org/papers/w20916 ↩ ↩²

18. Brand Science Institute — "If people want to find you, they'll find you." ↩

19. Jim Mollica, former VP Consumer Marketing at Bose — "Are we giving the cash register credit for the sale?" ↩

20. Viant/ANA — 70% of ad budgets to demand capture, 30% to demand creation. ANA, 2023. ↩

21. Ahrefs AI Overview citation analysis — top 10 organic citations dropped from 76% to 38%, early 2025 to March 2026. See: ahrefs.com/blog ↩

22. SE Ranking — 30,000 commercial keywords, January 2026. Top 5 review platforms: 88% of citations, 76-92% traffic loss. G2 -84.5%, Capterra -89%, TrustRadius -92.2%. See: seranking.com/blog ↩

23. Alphabet Q4 2025 earnings — Google Search revenue $63.07 billion, up 17% YoY. See: abc.xyz/investor ↩

24. Nick Fox, Google SVP/GM Search — February 2025 earnings call: "Google Search is becoming AI search." See: abc.xyz/investor ↩

25. US Patent 12536233B1, January 2026 — AI-generated landing pages hosted by Google with product information, pricing, and direct conversion paths. See: uspto.gov ↩

26. Semrush data — Reddit search visibility +1,274% year-over-year. See: semrush.com/blog ↩

27. Programmatic SEO outliers — Flyhomes +10,737% (location landing pages, via Search Engine Land 2024), Quillbot +490% (AI writing tools), CK12 (interactive educational content). Industry reports, 2024-2025. ↩

28. First Page Sage longitudinal B2B SaaS SEO study — 702% average ROI, 7-month break-even. See: firstpagesage.com/seo-roi ↩ ↩²

29. BrightLocal local search statistics — 76% of local searchers visit a business within 24 hours. See: brightlocal.com/learn ↩ ↩²

30. VCCP — Hacking the Attention Economy (Nelson-Field data). Top 50 advertisers: 3% revenue growth vs 4x digital ad spend. 66p of every £1 wasted on sub-threshold attention. Gen Z: 46% prefer social over search for product discovery; Instagram 30.4%, TikTok 23.2%, Google 18.8%. Email $36-$45 ROI. Podcast 500M listeners, 80% listen-through. YouTube 2.5B users, 70% mobile. See: https://pages.vccp.com/hubfs/Hacking%20the%20Attention%20Economy%20Report.PDF.pdf ↩ ↩² ↩³ ↩⁴