The honest traffic picture — what the numbers actually say
Google organic is still growing in absolute terms. Its share of total discovery is shrinking. Both things are true.
The 'is SEO dead' debate is usually conducted with rhetoric more than data. The honest picture is messier than either camp admits. Google organic search volume continues to grow year-over-year in absolute terms — Google reported 5 trillion searches in 2024, up from 3.5 trillion in 2020. At the same time, Google's share of all discovery has shrunk because AI answer engines have captured incremental queries, especially conversational and high-consideration ones. For a Shopify merchant, both facts matter: the absolute number of potential Google customers is not declining, but the share of the discovery pie you capture through Google alone is.
- Google search volume 2020
- 3.5 trillion queries/year (Google disclosed).
- Google search volume 2024
- 5 trillion queries/year — up ~43% despite 'SEO is dead' discourse.
- Shopify organic traffic share 2020
- 65-80% of total non-paid traffic for mature stores.
- Shopify organic traffic share 2026
- 40-60% for mature stores; rest split between AI engines (15-30%) and social (10-15%).
- ChatGPT weekly users 2026
- ~700M weekly active users globally (OpenAI disclosed, March 2026).
- Perplexity monthly queries 2026
- ~400M/month estimated, growing 8-12% monthly.
40-60%
of mature Shopify stores' non-paid traffic still comes from Google organic in 2026
Surfient attribution panel, 847 stores. Share has dropped from 65-80% in 2020 — but absolute Google-organic traffic for these stores is essentially flat, not declining.
What actually is dead — the old SEO playbook from 2015-2022
Thin content, keyword-stuffed articles, directory links, anchor-text over-optimisation, exact-match listicles. These tactics stopped working before AI search arrived.
The version of SEO that is genuinely dead is the low-effort, high-volume playbook that powered a lot of mid-market ecommerce content from 2015 through 2022. Content farms producing 500-word articles on every possible exact-match keyword variation. Mass directory submissions and link farms. Aggressive anchor-text optimisation. Listicles written by writers with no category expertise. Google's Helpful Content Update series (2022-2024) hit these hard, and the AI search layer piled on — retrievers read for substance, not keyword density.
Specific tactics that stopped working
- Thin article content (under 800 words) targeting exact-match long-tail keywords. Helpful Content Update devastated this pattern in 2022-2023.
- Mass backlink acquisition through guest posts, paid placements, and link exchanges. Google has filtered most of these at the index level.
- Exact-match domain and URL stuffing. Domains with keyword-stuffed TLDs underperform generic brand domains consistently.
- Generic 'top 10' listicles written without first-hand experience. E-E-A-T updates specifically target these.
- Anchor-text over-optimisation. Modern link graphs penalise over-optimised anchor distributions.
- Programmatic SEO pages at scale without substantive content. Worked in 2018, broke in 2023, irrecoverable in 2025.
What was always the actual work
- Clear site architecture and internal linking that matches user intent.
- Substantive content with genuine expertise and verifiable claims.
- Schema markup that accurately describes the page's content.
- Good technical hygiene (Core Web Vitals, HTTPS, sitemap, crawlability).
- Relationships with other sites earned through genuinely useful content.
What survived — the 60-70% of SEO practice that still moves the needle
Technical hygiene, schema, content quality, internal linking, and earned authority. These still matter for both SEO and GEO.
A large majority of the classic SEO practice survives — in some cases unchanged, in some cases renamed. The technical and structural work that an SEO team did well in 2015 is almost entirely still valuable in 2026. What has changed is the evaluation layer: the same work now has to satisfy both Google's algorithm and a half-dozen AI retrievers simultaneously, which raises the bar on quality and lowers the tolerance for gaming.
- Technical hygiene
- HTTPS, Core Web Vitals, crawlability, XML sitemap, clean URL structure, mobile-friendly rendering. 100% unchanged — required for SEO and GEO.
- Schema markup
- Product, FAQPage, Organization, BreadcrumbList, HowTo, Article. More important for GEO than for classic SEO today.
- Content quality
- Substantive, first-hand, expert-authored content. The bar has risen; the discipline is unchanged.
- Internal linking
- Topic clusters, pillar pages, meaningful anchor text. Works for both SEO and GEO.
- Earned authority
- Backlinks still matter (less than they did), mentions matter more, cross-source corroboration matters a lot more for GEO.
- Site architecture
- Logical taxonomy, clear hierarchy, user-intent-aligned URLs. Required for both.
The real shift — from page rank to passage retrieval
The architectural move underneath everything: search is moving from ranking pages to extracting passages. Your content needs to work both ways.
Underneath the 'SEO vs GEO' debate sits one architectural shift that explains most of what is changing: search is moving from ranking whole pages to extracting individual passages. A Google blue-link result ranks your page as a unit; an AI answer quotes a specific paragraph or sentence. That change cascades into almost every tactical difference merchants notice. It is why description openers matter more than they used to, why FAQ schema carries more weight, why short and self-contained content blocks outrank long run-on paragraphs. The page still matters; the passage matters more.
Content architecture implications
- 1Lead with the answer. The first sentence of any section should answer the question implied by the H2. Classic journalism discipline, now algorithmically enforced.
- 2Make every passage self-contained. A paragraph that relies on the previous paragraph for context cannot be quoted in isolation. Rewrite for independence.
- 3Use structured data generously. Lists, tables, key-value pairs, FAQ blocks, step lists — all of these give retrievers cleaner extraction targets than prose.
- 4Prune tangents and transitions. Connective prose that bridges sections is low-signal in a passage-retrieval world. Tighten or remove.
- 5Write H2s as questions. A question-shaped H2 matches prompt-shaped retrieval queries more directly than a statement-shaped H2.
“The shift from page rank to passage retrieval is the 'mobile-first' of our decade. Everyone will eventually do it. The stores that do it early will compound advantage over the stores that wait.”
Budget advice for a merchant rebalancing in 2026
Do not cut SEO. Do not ignore GEO. Fund GEO from new budget or from channels that are genuinely underperforming.
The most common failure mode we see on merchant-advisory calls is a panic cut of SEO spend to fund GEO. The merchant reads 'SEO is dying', reallocates 40% of the content budget to GEO work, watches Google organic decline 18% over the next two quarters, and realises they have cut what was still producing revenue to chase a channel they were going to build into anyway. The honest budget pattern is additive, not substitutive, at least for the first 12-18 months.
- Mature SEO program, AI exploratory
- Keep SEO spend flat. Add GEO at 25-35% of current SEO budget, funded from new allocation or from clearly-underperforming channels.
- Mature SEO program, AI growing share
- Keep SEO spend flat to slightly declining. Grow GEO to 50-70% of SEO spend by year-end 2026.
- Weak SEO program, AI growing share
- Ship SEO baseline (technical hygiene, schema). Spend heavier on GEO content. SEO here is prerequisite, not a program.
- Net-new store
- Start with combined SEO + GEO content work — the structural moves overlap by 70%. One well-built page satisfies both.
A three-year outlook — what 2029 probably looks like
Google AI Overviews grows, classic blue links shrink, ChatGPT-class engines consolidate at 30-40% of commerce discovery. Merchants who prepared win.
Forecasting three years out in search is inherently uncertain, but the trajectory signals are clear enough to plan against. Our working model for merchant advisors assumes Google AI Overviews continues to grow as a share of Google search pages, classic blue-link real estate continues to shrink, and ChatGPT-class engines (including Perplexity, Claude, Gemini, Copilot) consolidate at roughly 30-40% of commerce discovery by 2029. That leaves Google as the majority channel but no longer the dominant one, and it makes multi-channel visibility the default rather than the specialist position.
- Google AI Overviews probably covers 70-85% of commercial queries by 2029, up from 35-50% in April 2026.
- ChatGPT Shopping and its successors likely handle 15-25% of transactional commerce prompts directly, with checkout-in-chat becoming table stakes.
- Classic Google organic probably still drives 25-40% of Shopify traffic for mature stores — smaller share, similar absolute numbers.
- E-E-A-T signals continue to outweigh link signals for high-trust categories (health, finance, high-ticket commerce).
- Schema.org adoption continues to rise as the lingua franca of structured data across engines.
“Merchants asking 'is SEO dead' in 2029 will sound to us like merchants asking 'is mobile dead' in 2015. Wrong question. The right question is: where is the next quarter's marginal dollar earning the most?”