What is SEO?

SEO is the discipline of being chosen by every engine that mediates discovery — Google for the open web, ChatGPT and Gemini and Claude and Perplexity for AI-native search, AI Overviews for the result page itself. The work is the same shape it has always been: make a page that genuinely answers a query, prove to the engine that the page is trustworthy, and get the architectural details right so the engine can read it. What changed in 2026 is the number of engines that have to pick you.

The popular definition — “optimizing your website to rank higher on Google” — is not wrong, just thirty percent of the actual job. This article is the full one.

“Most articles on “what is SEO” explain how search engines worked in 2014. Treat the rest of this page as a reset.”

The 30-second definition

SEO (search engine optimization) is the practice of making a website legible, trustworthy, and useful enough that engines — both classic Google search and AI answer engines — pick it as a source when a relevant query comes in. It is a long-cycle discipline (cycles measured in weeks for indexation, months for ranking, sometimes years for category authority), and it compounds: every page that earns trust strengthens the next one.

Three things follow from that definition that the simpler version misses:

• Multiple engines, one signal stack.The technical, content, and authority signals that earn a Google ranking also earn a ChatGPT citation. SEO in 2026 is not “Google SEO + AI SEO” as separate practices — they share inputs.
• Picked, not crawled. Indexation is the floor. The actual mechanism is the engine choosing your page from the index when scoring a query against millions of candidates. Optimization is about being chosen, not just being there.
• Discipline, not tactic. SEO is a programme that runs every week, not a checklist you complete. Sites that treat it as a one-time deliverable lose ground inside six months.

What “optimization” actually means

The word “optimization” in SEO is misleading. It implies tweaking — adjusting a knob to push a number up. The real work is closer to product design: you decide what query the page should answer, structure the page so the answer is unambiguous, ship the technical foundation that lets engines read it, and accumulate the authority signals that prove the answer is credible.

Tweaks happen at the edges (a rewritten title, a tightened meta description, a fixed broken canonical), but the centre of the discipline is design: choosing the right page-to-query match in the first place. Sites that obsess over micro-tweaks while shipping the wrong content rarely move; sites that get the page-to-query design right and only do moderate hygiene afterwards usually win.

How SEO works in three layers

Modern SEO is best modelled as three layers that have to work together. Pages that succeed are good across all three. Pages that fail usually fail because one layer is broken — and the broken layer is rarely the one the team has been working on.

The technical layer

If technical is broken, nothing else matters — the engine will not see your page or will misread it. The 2026 non-negotiables: Core Web Vitals passing on mobile (LCP under 2.5s, INP under 200ms, CLS under 0.1), JSON-LD schema markup on every template, an XML sitemap segmented by content type, HTTPS, and a robots.txt that reflects your actual indexation intent. Most of this is shippable in a week by a competent developer.

The content layer

This is where the work compounds or evaporates. Each page targets a query a real user types or asks; it answers that query in the first 150 words; it carries the structural cues (clear headings, FAQs, comparison tables) that both Google and AI engines parse natively. The mistake is producing content for keywords without ever answering a query — generating volume without earning a ranking, then blaming the algorithm.

The authority layer

The signal engines use to decide whether your page is trustworthy. Backlinks are the classic version (still important); brand mentions in editorial coverage, podcast quotes, conference talks, and Wikipedia citations are the 2026 version that AI engines weight heavily. Without this layer, technically perfect content sits at position 30 forever.

Why SEO is bigger in 2026 than it was in 2018

The popular narrative — “AI killed SEO” — has the direction backwards. SEO is bigger now than at any point in its history because the number of surfaces that mediate discovery has grown.

In 2018, “ranking on Google” was the entire game. In 2026, the game is being chosen across a portfolio of surfaces:

• Google organic SERP — still the largest discovery surface by query volume.
• Google AI Overview — the AI-generated summary above organic results, citing 4-8 sources per answer.
• ChatGPT — primary AI answer engine for many B2B and consumer queries; cites 3-6 sources via web search.
• Gemini — built into Google products, with its own citation pattern.
• Perplexity — answer-engine-first interface, cites every claim.
• Claude — added web search and citation in 2025; growing fast in technical/research queries.

All six surfaces use overlapping signals to decide who gets cited. SEO is the discipline of optimizing for that overlap.

What SEO is not

Four misconceptions that still cost teams budget every year. Naming them so you can spot them in your own thinking and in vendor pitches:

• SEO is not link buying. Paid link networks have been a manual-action trigger for over a decade. Authority comes from earned mentions, not purchased PageRank.
• SEO is not keyword stuffing. Repeating a phrase 20 times in a paragraph is a 2008 tactic that triggers spam classifiers in 2026. Modern keyword targeting is one canonical mention per critical structural element (H1, title, first paragraph, URL slug) plus natural usage in the body.
• SEO is not a one-time project. Ranking is a continuous score that updates with every Google index refresh and every AI engine retrain. A site that stops investing loses ground because competitors keep moving.
• SEO is not separable from content marketing.“The SEO team handles technical, the content team handles articles” is the org structure that produces good articles nobody finds. Modern SEO and content are the same discipline operating on the same artefact.

When SEO is the right channel — and when it isn’t

SEO is the highest-leverage channel for businesses where (a) prospects have informational queries before they buy, (b) the buying cycle is longer than a click, and (c) competitors are also visible in search. That covers most B2B SaaS, professional services, healthcare, education, ecommerce above $50 AOV, and local services with map-pack relevance.

SEO is a poor primary channel for: pure impulse-buy products under $20 where no informational query precedes the purchase; brand-new categories nobody is searching for yet (build paid + community first, then SEO once a query forms); and businesses with sub-30-day buying cycles where the channel cannot ramp before the customer chooses someone else.

For everyone in the first list, SEO compounds: each piece of content keeps producing pipeline for years, the technical foundation pays back across every future ship, and the authority signals strengthen every adjacent ranking.