Found

We ran 100 unbranded travel searches through Google AI Mode and logged every source. It built its answers from 2,558 citations across 1,229 domains, about 26 per answer, and 75% of everything it cited was a guide, a tips piece, or an itinerary, the exact content most teams are quietly cutting. This episode walks through why AI Mode harvests rather than ranks, why the content everyone is trimming i...

Ask ChatGPT the same question twice and you often get two different answers citing two different sets of sources. Ask Gemini, Perplexity, and Google AI Overviews and the gap widens. That means turning up in one answer proves almost nothing, and one strong channel is not enough. This episode explains Total Graph Authority, the position a brand holds when three independent sources of trust all point...

When an Indonesian asks AI Mode about a mortgage, a savings account, or a credit card, someone's website is being quoted back as the answer. We measured whose. This episode walks through a study of 72 Bahasa Indonesia prompts run three times each through the Google AI Mode API, 216 answers and 2,068 citations across 312 domains. We cover why banks own most of the answer (close to 62% first-party o...

For about a decade, the keyword cluster was the atom of content planning. But AI Mode queries run roughly three times longer than old searches, and the engine no longer matches a string to ten blue links, it breaks the question into sub-questions, retrieves passages, and assembles an answer. This episode covers the shift to planning in prompt clusters instead: what keyword clusters were good at, w...

A dofollow link from a strong site used to be the most valuable thing you could earn off your own pages. In 2026 there is a real case that a plain, unlinked mention of your brand name does more, at least on the surfaces where buyers now start their research. This episode separates the two signals: why backlinks still move Google rankings while brand mentions are what get you named inside an AI ans...

We ran 105 buying-intent travel queries in Bahasa Indonesia through Google's AI Overviews across two days and logged every citation. Instagram was cited 104 times, more than Traveloka, tiket.com , or any other travel platform, and reels were quoted as evidence for prices and packages, not just as profile links. This episode walks through the full dataset: how Instagram works as citation infrastruc...

68% of US Google searches now end without a click, up 7.5 points in two years, the fastest acceleration on record. Counted in raw clicks, the open web lost roughly a quarter of its search clicks in 24 months while Google's ad revenue kept growing. This episode reads the 2026 SparkToro data closely: what is driving the jump (AI Overviews, re-queries, and a still-tiny but fast-growing AI Mode), wher...

GEO so far has been a citation game, getting named in an answer a human reads. A second game has started on top of it, and this one has no human reading anything. This episode covers what agentic AI does to generative engine optimization: how an agent scoring vendors queries your structured data and drops anything it cannot parse, why fixing the schema markup you already have comes before adopting...

One browser API can drop p75 LCP from 1,800ms to 320ms with no image work, no CSS surgery, and no server migration, and Google Search itself started using it to prefetch results before you click. This episode is a practical guide to the Speculation Rules API for SEO teams: why your field data moves in Search Console even when lab scores do not, which URL patterns are worth prerendering versus pref...

Search no longer hands people ten blue links, it reads the web, synthesizes an answer, and puts a machine between you and your reader. This episode covers how to publish helpful, people-first content that still gets cited by AI Overviews, ChatGPT, and Perplexity. We break down what the March 2026 core update actually rewarded, why SEO, GEO, and AEO are three jobs rather than three names for one, h...

Every few months someone declares SEO dead, and the numbers keep refusing to agree. The field is not dying, it is splitting in two, and the growing half looks almost nothing like the work that defined it five years ago. This episode covers how AI search rewrote which skills get you hired: the entry-level tasks the machines already took, the four new workstreams that replaced the old org chart (AI...

Pull any large site's crawl logs and a real slice of bot activity is spent on URLs that died months ago. This episode cuts through the 410 versus 404 myth: the deindex-speed gap is small, but the recrawl-frequency gap is large and measurable, with one controlled experiment finding 404s crawled nearly 50% more often than 410s. We cover what each status code actually tells a crawler, why this is a B...

A buyer asks ChatGPT whether your product is any good, and the answer comes back wrong: an invented founding date, a service you never sold, or a complaint a competitor seeded months ago. You never see the query, and the deal quietly disappears. This episode covers defensive GEO, the blind spot in a field obsessed with offense. We break down the two threats that wear the same mask (hallucination v...

IndexNow gets dismissed for one reason: Google does not support it. But Google is not the only index that matters anymore. This episode explains why IndexNow is really a freshness pipe into the AI answer engines, since Bing's index feeds Microsoft Copilot, DuckDuckGo, and the web search behind ChatGPT. We walk through the full setup end to end, generating a key, hosting the key file, and pinging o...

Lovable, Bolt, v0, Replit, and Cursor can ship a working site before lunch. The problem shows up weeks later: most of them ship a client-rendered app that ChatGPT, Claude, and Perplexity read as a blank page. This episode breaks down why that happens, how each builder handles rendering today, why Bing quietly decides whether ChatGPT can find you, and the exact fixes and pre-launch checklist to get...