Naver has never been known as a simple search engine. As pointed out in my recent article on Naver’s paid search, it's a dynamic, complex software system appealing to an affluent yet small market by Asian standards (South Korea’s population is only approximately 50 million). Considering its challenges, paid search on Naver at least guarantees you a top spot, but, what about organic?
Naver’s search results are categorically grouped. Its organic results are based upon three different areas of content:
A question and answer (Q&A) knowledge base (Knowledge Q&A) A database search (Direct Answers) Websearch (organic search results), which are within one massive result page
Why? According to Naver, Korean web content is miniscule compared to other languages like English and Spanish.
Naver puts in a lot of effort to provide more enhanced content rich search results to users in a way that appeals to the Koreans need for as much information as possible. It's a supply and demand thing really, when you have “less supply” (Korean web content), yet high demand (search queries), you’re forced to provide as much quality content as possible. In order to do this, Naver figured out that it might as well provide three separate ways to provide all of this information.
“Home Loans” keyword search on Naver:
Considering the lack of enough “web content” in Korean, Naver figured out that user-generated content (UGC) might be the key to solving its lack of content problems. So in 2002, they were one of the first search engines to create a Web 2.0 knowledge base Q&A portal where users could pose questions and others could answer. In 2005, Yahoo based its U.S. Q&A site on the same model. Still today, its one of the largest traffic referring sources from (ironically) Google.
Posing related questions to your business and having people answer those questions, will help your positioning in the Q&A section. Keep in mind, popularity is key here, so you better have your story straight or you run the risk of embarrassment.
Database Search (a.k.a. Direct Answers)
Another way Naver has been able to build up content on their portal is by incorporating what they call a “database search” from a vast archive of content partners. The content partners feed their database content by allowing Naver to crawl and index their results within their search result pages. These database searches are common for celebrities, weather, movies, books, wines, public services, famous brands, etc.
There is a fine line when it comes to organic database search versus paid database search. For example, a Wikipedia type site feeding its useful content into Naver would be considered non-paid, because it’s mainly content rich value.
However, a massive site and brand like Intel could also plug in their databases of products into Naver, but for a fee. This is similar to paid inclusion, but Brand Search results in awesome, top of page and 60 percent viewable real estate in the results.
Finally, when you reach Websearch results, you have reached the natural, organic search results as we know it in the SEO industry. There are a number of similarities to Google however, weighting factors are completely different. Naver has two methods to analyzing and ranking, one known as “collection ranking” and the other is known as “search consequence ranking.”
- Collection Ranking: Collection ranking is basically impression based click-through rate (CTR) analysis ranking and is based on an average click index that occurred from classified collection sets (similar to ad groups).
- Search Consequence Ranking: There are many other factors for the rest of their algorithm, such as keyword relevance, website content freshness, content accuracy, quality of information, quantity of information, and the inherent link popularity of the page. Naver completely excludes indexing low quality pages, spam and commercial pages, they purely want search results that have what they consider great content. Using what they call “FAS” (Feedbank Analysis System), Naver reflects the search ranking by the duration of user’s visit and “collection ranking” to determine the quality of the link pages.
Although, Naver provides organic results, the best way to do it is with tons of good quality content. The investment in that amount of content versus the investment in paid search is to be weighted.
Similar to Yandex and how it understands the Russian language through semantic intent mixed with its language analysis, Naver has also built its engine around the language. Naver developed its indexing engine for the Korean keywords and language, along with categories, resulting in a better user experience and better results then its competitors.
Due to the nature of Korean language syllabification, it’s harder to understand the meanings by separation than other languages. For example, “ilbonjeon gol” in Korean means “goal at the match with Japan”; however, “ilbon jeongol” means “Japanese stew.”
Naver: Complicated, Overwhelming...Awesome
I never said Naver would be easy. But considering Google’s near zero market share, Naver is doing something right.
Naver feels that users can’t be perfectly satisfied with their search intention (a Google experience). Hence, Naver offers the search results of the various areas and categories such as news, Knowledge Q&A, blogs, paid, and data base searches all on the same page. Each individual category has its own ranking methodology.
To our eyes, it looks complicated and overwhelming. To a Korean, it’s “awesome.”
Huge thanks to Kyle Choi of NHN (Naver’s U.S. office) for helping me figure this out.