Category-keyword domains, versus everything else.

For most of the post-search-engine era, the conventional answer to "where should a category brand live online" has been a sub-path under a larger parent property. Marriott runs marriott.com/destinations/korea. Trip.com runs trip.com/travel/korea. Lonely Planet runs lonelyplanet.com/south-korea. Each of these is a sensible architectural choice from inside the parent organization. From outside, from the perspective of a user trying to find Korean travel information by typing into a browser, none of them is the obvious destination.

This essay walks through why category-keyword domains continue to outperform sub-paths and sub-domains for brand discovery in 2026, sourced to the traffic data, the AI-search behavior, and the operator economics. The implication for any operator considering a Korean-content, Korean-tourism, or Korea-facing commerce position is structural rather than tactical.

The traffic data, in plain terms

Direct-navigation traffic, broadly defined as visits where no referrer is recorded, sits between 22 and 28 percent of total web traffic in published 2025-2026 industry data.^[1] What varies is which sites accumulate disproportionate shares of that channel. Public Semrush data for Domain.com shows roughly 70 percent of its visits arrive via direct navigation.^[2] The pattern is consistent across category-keyword domains: the more distinctive and category-defining the name, the higher the share of arrivals via the address bar.

Booking.com is the textbook case. The site accumulated meaningful direct-navigation share in the late 2000s and 2010s through aggressive paid acquisition, then compounded that direct-traffic position through the next decade as the brand became synonymous with hotel booking. By 2025, Booking.com sits among the highest-direct-traffic sites in the global travel category, with the address itself functioning as a primary discovery surface alongside search engines.

Hotels.com operates the same pattern at a smaller scale, and Expedia Group's portfolio across multiple category-keyword domains (Hotels.com, Vrbo.com, Trivago.com) collectively captures attention that no single brand operating from a sub-path could replicate. Our earlier essay on direct-navigation traffic covers the broader mechanics; this essay focuses specifically on the structural reasons category-keyword domains beat sub-paths and sub-domains for the same job.

Reference traffic concentrations, 2025-2026.

Domain.com: roughly 70 percent direct (Semrush, Nov 2025)
Premiumy.net: nearly 90 percent direct (Semrush, mid 2025)
All-site average: 22 to 28 percent direct (WifiTalents Feb 2026 and Sojern 2026)
Long-form content sites with SEO focus: typically 10 to 20 percent direct
Country-keyword domains in mature operating businesses: typically above 40 percent direct

Why sub-paths underperform on the same job

A sub-path under a parent property (parent.com/category) and a sub-domain (category.parent.com) both inherit some authority from the parent domain but pay structural costs that a standalone category-keyword domain does not. Four observable disadvantages.

One: address-bar instinct does not produce sub-paths. A user who has just finished a Korean drama and wants to find more does not type marriott.com/destinations/korea into a browser. They do not type any sub-path. They type a category-keyword. Whether that user lands at any specific destination is a function of which operator owns the category-keyword address; the sub-path is invisible to the direct-navigation channel by construction.

Two: bookmark behavior favors short, memorable addresses. The set of addresses a user types repeatedly converges over time on the few that are short enough to remember and distinctive enough to keep. Sub-paths almost never make that set. Sub-domains rarely do. Category-keyword domains, if they are short and memorable, are the natural occupants of the position. The compounding traffic effect of being in that set produces the kind of direct-share numbers visible at Domain.com and Booking.com.

Three: word-of-mouth transmission strips the sub-path. "Go to korea.tv" survives transmission cleanly in conversation, in social media, in voice search, and in casual reference. "Go to marriott.com/destinations/korea" does not survive any of those transmission paths. The sub-path becomes a search query, which dilutes attribution and pushes the user through a search engine that may route them to a competitor.

Four: AI search recommendations cite addresses, not sub-paths. Through the 2025-2026 generative-AI-assistant era, an emerging pattern is that AI assistants increasingly cite and link to specific addresses in their answers. Users see the address, recognize the brand, and the next time they want the same kind of answer, they often bypass the assistant and go directly to the address. The shorter and more memorable the cited address, the more often the user returns directly. Sub-paths are cited less often by AI assistants and remembered less reliably even when cited.^[3]

The compounding effect, over time

The structural reason category-keyword domains accumulate value rather than depreciating is that direct-navigation traffic is the only acquisition channel that gets cheaper over time. Search-engine traffic costs more every year as competition for keywords increases. Paid advertising costs increase as platforms mature and ad inventory tightens. Influencer and social channels have flatlined or declined in efficiency for many categories. Direct navigation, by contrast, is a free channel whose cost-per-visit asymptotes toward zero as the underlying name's recognition grows.^[4]

The compounding works because each direct visit increases the probability of another. A user who types your address once is more likely to type it again. A user who recommends your address verbally adds new direct visitors to the pool. A user who bookmarks your address makes the next visit faster. Over years, this compound creates a traffic position that no marketing spend can economically replicate.

This is the reason category-keyword domains hold their value through every shift in the underlying internet. The underlying technology changes. The dominant referral channels change. The address bar does not change.

What this means for country-keyword domains specifically

Country names are exceptionally well-positioned for direct navigation. A country name is, by definition, one of the few addresses a user can guess without prompting. A user curious about Korean culture, Korean television, Korean tourism, Korean food, or Korean music will, in some measurable fraction of cases, type the country name plus a category-suffix extension as their first attempt. The same logic applies to other geographic-keyword properties.

Three structural features make country-keyword addresses more durable than even category-keyword addresses.

One: the universe is small and fixed. There are roughly 200 countries with internationally recognized names. The set does not grow. Owners of geographic-keyword properties in the major extensions know this and price accordingly.

Two: the use case is multi-vertical. A country-keyword address is plausibly the right address for a tourism brand, a media brand, an export brand, a cultural-content platform, and several other categories at once. That breadth makes the asset more valuable to an operator with a plan than to a passive investor. Our essay on the Korea brand in 2026 covers the four-vector commercial pull in detail.

Three: the brand fact is irreplaceable. A country brand is one of the few brands that no marketing spend can construct. It is built by the country itself, over decades, through cultural exports, tourism, diaspora communities, and global media presence. Korea's brand fact is what makes a Korea-keyword address valuable; the address simply captures attention that the underlying country has already built.

The architecture question for operators

For any operator considering a Korea-facing brand, the architectural choice between a category-keyword domain and a sub-path under a parent property is not a styling decision. It is a structural decision that shapes the next decade of acquisition economics.

A sub-path under a parent platform produces traffic only to the extent that the parent platform produces traffic. The sub-path inherits parent-domain authority for search but does not capture any direct-navigation share of its own. The unit economics work as long as the parent traffic continues to flow.

A category-keyword or country-keyword anchor produces traffic that compounds over time, independent of any single parent platform's performance. The unit economics improve every year. Acquisition costs fall. Brand-related trust signals accumulate. And the architectural moat is structural: the address itself is unique, and no late entrant can replicate it.

The trade-off is capital. A sub-path costs nothing additional to operate. A category-keyword anchor requires acquiring the address. The economic question is whether the multi-year compounding value of the anchor exceeds the one-time acquisition cost. For a category as large and durable as Korean cultural exports, the answer is almost always yes for operators with the capital and the plan to execute on it.

For broader context on the foundational case: essay one covers the underlying Korean cultural-content economics, essay two covers the cross-vector brand pull, essay three covers the global audience side. For the acquisition mechanics: essay four covers discovery, essay five covers closing, essay six covers the channel mechanics. For the operator-side strategy: essay seven covers tourism, essay eight covers streaming, essay nine covers commerce. For the most recent strategic moves: essay ten covers global streaming-platform architecture, essay eleven covers M&A activity, essay twelve covers DMO digital portfolios, essay thirteen covers the K-content aggregator map, and essay fourteen covers the Korean broadcaster international expansion plans.