Search Marketing FAQ

It's when two or more pages on your site target the same keyword, splitting authority and confusing Google about which to rank. The fix is content consolidation: merge the pages, redirect the weaker into the stronger, and rebuild internal links to point at the surviving URL.

in SEO Program

Search intent is the goal behind a query - informational (how does X work), navigational (X login), commercial (X reviews), or transactional (buy X). Ranking requires matching the intent: a product page won't rank for an informational query no matter how strong the SEO, because Google has decided users want articles. We classify every priority keyword by intent and align the page type accordingly.

in SEO Program

Use them where they read naturally - page title, H1, opening paragraph, URL slug - and stop. Modern search algorithms (BERT, MUM) understand semantic meaning, so keyword stuffing is both unnecessary and a quality signal flag. The goal is topical depth across related entities, not raw keyword density.

in SEO Program

Internal linking is how pages on your site reference each other. It distributes authority (PageRank flow), signals topical relationships, and helps Google understand site hierarchy. A page with strong external links but no internal links is wasting authority; a page with strong internal links but no external is starved for authority. You need both.

in SEO Program

A 404 is the response code when a URL doesn't exist. Occasional 404s are normal. The problem is when 404s replace pages that previously ranked or had backlinks - that's lost equity. The fix is auditing 404s monthly and 301-redirecting any that had organic value to the most relevant live page.

in SEO Program

A canonical tag tells Google this is the preferred URL for this page. It's how you handle near-duplicate pages (filtered listings, tracking parameters, paginated archives) without losing authority. Misconfigured canonicals are one of the most common SEO leaks we find on audit.

in SEO Program

With a migration playbook: full URL inventory pre-launch, redirect map (every old URL to its new equivalent), schema and metadata parity check, internal-link rewrite, sitemap re-submission, and a 30-day rankings monitor. Migrations done without this lose 20-60% of organic traffic; migrations done with it preserve 95%+.

in SEO Program

They don't hurt, but they don't help - subdomains are treated as separate sites by Google for the most part, which means authority doesn't flow as freely between blog.example.com and example.com as it would between example.com/blog. For most marketing sites, subfolders are the higher-leverage choice.

in SEO Program

Structured data (JSON-LD schema) is machine-readable metadata that tells search engines what a page is about - Article, Product, Organization, Review, Event, FAQPage, BreadcrumbList, HowTo. Schema doesn't directly raise rankings but it earns rich results (star ratings, FAQ accordions, image carousels) that increase click-through rate, and it's a foundational signal for AI channel extraction.

in SEO Program

Search Console is the source of truth for how Google sees your site - impressions, clicks, average position, indexing status, crawl errors. Analytics is the source of truth for what users do once they arrive - sessions, conversions, behavior. SEO programs need both; using one without the other leaves the picture half-blind.

in SEO Program

GEO is the discipline of being the source an AI model cites or recommends when a user asks a question your business answers. The goal is presence inside ChatGPT, Claude, Perplexity, Gemini, and Copilot responses - not as a link buried in a list, but as the named source the model trusts.

in GEO Program

SEO optimizes for ranked lists; GEO optimizes for synthesized answers. SEO measures position; GEO measures citation. SEO weights backlinks heavily; GEO weights editorial mentions, primary-source authority, and content extractability. Mechanically there's overlap (both reward clean schema and quality content), but the optimization targets and reporting are distinct.

in GEO Program

Both, depending on intent. Informational queries often resolve in-chat (no click). Commercial-intent queries - comparison, pricing, vendor evaluation - produce high click-through rates because users want to verify before they buy. The traffic GEO produces is small in volume relative to Google but disproportionately high-intent; we see conversion rates 2-4x organic Google averages.

in GEO Program

Two paths. (1) Training: the model ingests web content during its training cycle (typically refreshed every 3-6 months). Mentions on authoritative sites - editorial roundups, Wikipedia, industry directories, publications - get baked into the model's weights. (2) Live retrieval: when a user asks a question, the model fetches real-time results (ChatGPT via Bing + SearchGPT, Perplexity via its own crawler, Gemini via Google). GEO targets both layers.

in GEO Program

A citation is when an AI model names your domain, brand, or content as a source in its answer. Citations show up as inline links (Perplexity, ChatGPT Search), as named brand mentions (closed-book ChatGPT recommendations), or as recommended-action links (Copilot suggesting your service). All three are tracked in our GEO reports.

in GEO Program

By active-user volume: ChatGPT (900M weekly), Gemini (integrated into Google AI Overviews and Workspace), Copilot (Microsoft 365 default), Perplexity (high-intent research users), Claude (Anthropic - heavy in technical and enterprise audiences). We prioritize based on which platforms your buyers actually use, which the audit surfaces.

in GEO Program

Share of voice is the percentage of relevant queries in which an LLM names your brand vs. a competitor. We run a monthly query panel - a fixed set of representative buyer questions - across each LLM and track presence, sentiment, and competitive context. It's the closest GEO equivalent to organic ranking share.

in GEO Program

The economics are unclear and changing. Some platforms (Perplexity Pro, ChatGPT for Business) experiment with sponsored placements; others stay editorial. We treat GEO as we treat organic SEO - earned, not paid - and monitor each platform's commercial roadmap. If a platform turns paid, we shift the optimization tactic accordingly.

in GEO Program

Often, but not always. AI rerankers add their own preferences on top of ranking - they favor pages with direct-answer leads, FAQ schema, primary-source data, and clean structural markup. A page can be page-1 on Google and still lose the citation to a page-3 result with cleaner extraction. GEO closes that gap.

in GEO Program

Answer extraction is what an LLM does when it pulls a specific sentence or paragraph from a page to compose its response. Pages that put the answer in the first 100 words, in a clearly delimited block (heading + direct answer), get extracted at much higher rates than pages that bury the answer in narrative. Restructuring for extractability is the single highest-leverage GEO content tactic.

in GEO Program

Two requirements. (1) Be widely mentioned across the open web in the right context - editorial roundups, best X articles, industry awards, Wikipedia. (2) Be retrievable by ChatGPT's live search layer for the underlying query. Closed-book recommendations come from training mentions; live-retrieval recommendations come from current rankings. GEO works both.

in GEO Program

Yes, with caveats. AI referrers show up in analytics - chatgpt.com, perplexity.ai, gemini.google.com, copilot.microsoft.com - but volume is often understated because of stripped referrers and in-app browsers. We supplement with citation tracking (which queries name your domain) to get the full picture.

in GEO Program

A primary source is original research, original data, or first-hand expertise. LLM rerankers strongly prefer primary sources over synthesized roundups because they're more defensible and less likely to hallucinate. A 1,200-word article with original benchmark data outperforms a 4,000-word roundup of other people's research, almost universally.

in GEO Program

AI models occasionally state false facts about brands - wrong pricing, wrong features, outdated leadership. The fix is to flood the open web with authoritative correct sources (your site, press releases, industry coverage), strengthen your Wikipedia-style entity presence, and submit corrections via the platforms that accept them (Bing Webmaster, OpenAI feedback). Hallucinations decay when the model has more correct sources than incorrect.

in GEO Program