Overseas Website Glossary: SEO, GEO, CDN, CMS, Hreflang, and UTM

The short version #

This isn't a Wikipedia dump. It's the list of terms our clients keep mixing up in the same meeting. The founder nods at "SEO." Marketing says "GEO." The contractor hears "hreflang" and "canonical" used like they mean the same thing, and twenty minutes in, nobody is changing the same thing in the same place. So we wrote down the fourteen terms that come up most in an overseas website kickoff: a one-line definition, a real example, and the trade-off you actually have to make.

If you're scoping a new site, drop this page into the PRD as the shared vocabulary. If you've already launched, run down the list and ask which one was skipped. Most of the projects we audit failed on definition alignment, not on engineering.

Quick groups #

Fourteen terms, four groups:

• Search visibility: SEO, GEO, sitemap, robots, canonical, schema
• Site platform and languages: CMS, hreflang
• Infrastructure: CDN, DNS, SSL
• Conversion and data: UTM, CTA, CRM

Read whichever group matches your current sprint. You do not need to read the whole thing in one sitting.

SEO #

Search Engine Optimization. Getting Google and Bing to rank your pages on page one when buyers search your category. The work splits into three buckets: technical (page speed, structured markup, sitemap), content (titles, copy, keyword coverage), and external signals (backlinks, brand mentions).

In business: a buyer in Düsseldorf types "industrial valve manufacturer china." Are you on page one? If not, no inquiry.

The trade-off: you don't need 100 SEO tasks at launch. You need the skeleton: title, description, H1, sitemap, Search Console, internal links. Google's SEO Starter Guide lays out the same baseline. Everything else is two to six months of post-launch work. See also SEO for Export Companies and Technical SEO Baseline.

GEO #

Generative Engine Optimization. Getting ChatGPT, Perplexity, and Google AI Overviews to cite your page when they answer a buyer's question. AI doesn't return ten links. It writes a paragraph and may cite a few sources. Your job is to write pages an AI can quote: clear questions, clear answers, verifiable numbers, entity signals like author and publish date.

In business: buyers no longer scroll to page two. They ask the AI, "Which Chinese manufacturers do small-batch valve customization?" The AI names three. If you're not one of them, you're invisible.

The trade-off: GEO isn't a separate engineering project. It sits one layer above SEO. Answer paragraphs, FAQ schema, clear entity pages. See What Is GEO and How Is It Different from SEO?.

CMS #

Content Management System. The admin interface that lets non-developers update pages, publish articles, and swap images. WordPress, Strapi, Contentful, Webflow are all CMSes. The opposite of a CMS site isn't a "template site." It's a site where every word change requires a developer ticket.

In business: marketing wants to add a new case study. They log in, write it, hit publish, and it's live in twenty minutes. No ticket, no sprint.

The trade-off: about 85% of small-to-mid-sized exporters are well served by WordPress plus a clean theme and eight to ten core plugins. Deep ERP/CRM integration or complex interactivity pushes you toward a headless CMS plus Next.js or Nuxt. See WordPress Overseas Website Architecture.

CDN #

Content Delivery Network. A global network of edge servers that cache your static assets (images, CSS, JS) so each visitor downloads from the nearest node. Cloudflare, Fastly, and CloudFront are the common ones.

In business: your origin is in Singapore. A buyer in Frankfurt sees first paint drop from 4.8 to 1.2 seconds once a CDN is in front. That difference moves both conversion and ranking.

The trade-off: an overseas site without a CDN is running naked. Cloudflare's free tier covers most exporters. If you also need to serve mainland China, set up a separate domestic delivery path. See Website Performance Across Regions.

DNS #

Domain Name System. The lookup that translates yourbrand.com into a server IP. After registering, you configure A records (point to the host), MX records (point to mail), and TXT records (SPF, DKIM, DMARC, the email auth that keeps you out of spam folders).

In business: a misconfigured DNS means either the site doesn't load or your quote emails land in junk. The second one is invisible to you. The buyer just never replies, and you assume they lost interest.

The trade-off: Cloudflare DNS or your host's managed DNS both work. The non-negotiable is getting all three email auth records right. See Domain, SSL, Email, and Analytics Setup.

SSL #

Secure Sockets Layer. The certificate that turns http:// into https:// and shows the lock in the browser. It's been TLS for years technically, but the industry still says SSL.

In business: a buyer who sees "Not Secure" closes the tab. Google also treats HTTPS as a ranking signal.

The trade-off: Let's Encrypt issues free certificates and most managed hosts auto-renew. Never let one expire. The moment it does, the entire site is unreachable. Set a calendar reminder, or use a host that handles renewal automatically. See Domain, SSL, Email, and Analytics Setup for the field-level checklist.

hreflang #

Language and region tag. Tells search engines, "this page is the English version for the US, that page is the German version for Germany." Lives in the <head>. Every translated set points at every other version, including itself.

In business: you have an English site and a German site. Without hreflang, a buyer in Munich searching your brand may land on the English page, and the German version may never get indexed at all.

The trade-off: URL paths (/en/, /de/) are simpler than subdomains. Don't skip self-references. Set x-default to the English version. And no, hreflang doesn't fix machine-translated copy. If the German page reads like Google Translate, German buyers still bounce. See Multilingual Site Structure and Hreflang.

canonical #

The "official URL" tag. Tells search engines which version of a page is the master when the same content sits behind multiple URLs (parameters, UTM tags, session IDs).

In business: your case study is reachable at /case/abc and /case/abc?utm_source=linkedin. Without a canonical, Google sees two pages and splits link equity between them.

The trade-off: every page should self-canonical. Pagination, filters, and UTM-tagged URLs all canonical back to the clean version. Canonical and hreflang don't conflict. Use both.

schema #

Structured data / Schema.org markup. A JSON-LD block in the HTML that tells search engines, "this is an article," "this is a company," "this is a FAQ." Powers rich snippets in Google and gives AI a clean view of what the page is about.

In business: a FAQ page with FAQPage schema gets the answer expanded directly in the search result; click-through rises. A company page with Organization schema gives AI Overviews the right name and logo to cite.

The trade-off: Service for service pages, Article for posts, Organization for the company page, FAQPage for FAQs. Don't stuff in unrelated types. Google penalizes schema abuse aggressively. See Schema Markup for Service Websites.

sitemap #

sitemap.xml. An XML file that lists every indexable page on the site. Submitted to Google Search Console and Bing Webmaster Tools so crawlers know what exists and what changed.

In business: a new service page usually gets indexed within two days of being added to the sitemap. Without one, indexing can take two to four weeks.

The trade-off: WordPress users let Yoast or Rank Math generate it automatically. Custom builds export a static one. Exclude drafts, archives, 404s, and pages that require login.

robots #

robots.txt. A plain text file at the site root telling crawlers which paths to skip.

In business: a developer copies Disallow: / from staging to production at launch. The whole site stays invisible to Google for two weeks before anyone notices Search Console is empty. We've seen this exact mistake twice this year.

The trade-off: open https://yourbrand.com/robots.txt in a browser before calling launch done. Admin paths, carts, and internal search results can be disallowed. The main site cannot.

UTM #

Urchin Tracking Module / source parameters. Query parameters on a URL, like ?utm_source=linkedin&utm_medium=post&utm_campaign=launch, so Google Analytics can attribute the visit. Five standard fields: source, medium, campaign, term, content.

In business: every link you publish on LinkedIn, X, email signatures, and WhatsApp short links carries a UTM. End of month, you can pull a report showing which channel produced which inquiry. Without UTMs, everything goes to "direct" and you have no signal.

The trade-off: agree on naming before launch. linkedin, Linkedin, and LinkedIn become three rows in the same report and turn the whole spreadsheet into noise. See UTM Tracking for WhatsApp, X, Forms, and Email Leads.

CTA #

Call to Action. A specific instructional line paired with a button or link, telling the visitor what to do next. "Get a Free Quote." "Book a 30-Minute Consultation." "Download the Spec Sheet."

In business: "Learn More" is a bad CTA. "Request samples (delivered in 7 days)" is a good one. Concrete, measurable, low cost to click.

The trade-off: cap the primary CTAs on a service page at two. One above the fold, one in the footer. More than that and visitors freeze.

CRM #

Customer Relationship Management. HubSpot, Salesforce, Zoho, Pipedrive. The system that consolidates form submissions, WhatsApp messages, and email leads into one place where every follow-up gets logged.

In business: a form lands in HubSpot, an owner is assigned automatically, and if there's no reply in three days the lead escalates to a manager. Without a CRM, leads scatter across three inboxes, two WhatsApp accounts, and a spreadsheet, and some always go missing.

The trade-off: HubSpot's free tier is plenty for year one. Move to a paid plan or to Salesforce when volume justifies it. See Inquiry Response Workflow.

Quick reference #

FAQ #

Do we need to know all thirteen before launch? #

Not all of them, but assign roles. The technical owner cares about DNS, SSL, CDN, sitemap, robots. The SEO owner cares about SEO, GEO, hreflang, canonical, schema. Sales and marketing care about UTM, CTA, CRM. The CMS belongs to everyone who edits content. When an unfamiliar word comes up in a meeting, do not nod through it — pull up this page.

Is GEO replacing SEO? #

No. SEO is the foundation, GEO is the layer above it. AI search is growing as a share of total search traffic, but classic search still drives the majority of B2B inquiries. Get the SEO skeleton right first, then add the answer paragraphs and entity signals GEO needs. See What Is GEO?.

Do hreflang and canonical conflict? #

They do not. Canonical handles "multiple URLs for the same language version of a page." Hreflang handles "different language versions of the same content." Use both together. Google Search Central's localized versions guide has the canonical example pattern.

We have no in-house SEO. Should we hire one? #

Not in year one. Get this glossary aligned with the dev team and marketing, run the Technical SEO Baseline checklist, and you'll cover 80% on your own. The remaining 20% (content depth, backlinks, AI visibility) buys cheaper as monthly support under our overseas website build and SEO/GEO support than as a full-time hire.

Get a diagnosis #

If you're scoping an overseas website, or you've launched but the team is using these terms inconsistently, bring your live pages, target markets, and primary services. We'll run this exact glossary against the site as part of a free initial review under our overseas website build and SEO/GEO support, and hand you a P0 fix list plus a three-month priority plan.