Naver does not publish a Google-style core-update calendar. What it does publish are ranking models for blogs and café posts, then product changes that move where those posts appear. The table is limited to launches with a company post, a press release, or a contemporary news report of a named system. A lot of Naver ranking movement never gets a name. That absence is part of the history.
| Date | Update | What changed |
|---|---|---|
| 26 June 2026 | AI Tab | Conversational search opened to all users on the Naver main page, mobile and PC, replacing the older Green Dot entry. |
| 27 March 2025 | AI Briefing | AI summaries on search, with shopping, places and short-form content in the same rollout. |
| 1 February 2024 | Popular Posts Smart Block | The VIEW area on integrated search becomes topic Smart Blocks. The VIEW tab splits into Blog and Café. |
| 11 November 2022 | FAQ and Q&A Smart Blocks | Knowledge iN, blog and café posts bundled into FAQ blocks (official sources) and Q&A blocks (everyday opinions), with AI snippets. |
| 18 November 2020 | D.I.A.+ | Deeper query-intent matching, pattern analysis and dynamic ranking on top of D.I.A., still paired with C-Rank. |
| 13 September 2018 | VIEW search | Blog and café results merged into a VIEW tab on mobile, ranked more by intent than by which Naver service published them. |
| 12 June 2018 | D.I.A. | Deep Intent Analysis: document scores for topic fit, experience, completeness, originality, freshness and abuse. |
| 29 January 2016 | C-Rank (Live Search beta) | Creator Rank in the Live Search beta: source trust from topic consistency, content quality and the engagement chain. |
| 29 November 2013 | SONAR | Original documents preferred over copies and scrapes. |
| 27 November 2012 | Libra | Quality-weighted blog ranking. First applied to slices such as restaurant queries, from early December 2012. |
Key algorithm events
For most of the 2000s Naver’s search was a portal problem, not a blue-link problem. Knowledge iN, blogs, cafés, news and shopping were separate collections. Public write-ups from 2006 to 2011 rarely name a ranking formula the way Yandex named cities or Baidu named weather systems. That decade is a genuine hole in this table. Anyone selling a “complete Naver algorithm timeline back to 2006” with a neat row every year is filling gaps.
The first named quality push that the Korean press actually covered is Libra, announced on 27 November 2012. Seoul Shinmun reported Naver’s line in plain terms: higher-quality results first, with the new blog logic starting on some keywords, famously restaurant searches, in early December. A year later, on 29 November 2013, Naver said original documents would be shown before copies. The Kyunghyang report of that day is the SONAR announcement. Together they are the pre-history of everything Naver later said about trust: reward the source that did the work, not the site that reprinted it.
C-Rank is the model working SEOs still have to live with. It became visible to the public on 29 January 2016, when Naver opened Live Search in beta and described a creator score built from context (does this blog stay on a topic), content (is the post substantial) and chain (do people actually use it). Contemporary coverage includes ZDNet Korea. Naver’s help centre still documents C-Rank as the blog-source model, without that launch date on the help page itself. The practical effect was harsh on keyword blogs that jumped topic every week. A new URL with a perfect keyword match lost to a creator who had been writing that subject for months. That is still the right mental model, even after the SERP chrome changed.
D.I.A. arrives on 12 June 2018, in Naver Search’s own blog: Deep Intent Analysis. Where C-Rank scores the source, D.I.A. scores the document against what people wanted for that keyword: topic fit, first-hand experience, how complete the information is, whether it is abusive relative to other posts, originality, and how time-sensitive the query is. Naver said the weights are re-learned and written back into ranking on a cycle. The same post is explicit that the two models are complementary, and that high C-Rank creators tended to produce high D.I.A. documents. If you only optimise one, you are flipping a coin.
VIEW is the interface change that made those models famous outside Korea. On 13 September 2018 Naver turned blog and café results into VIEW on mobile, so the user’s intent outranked the old “which tab did this come from” split. D.I.A.+ on 18 November 2020 is the upgrade Naver described on its search blog: deeper matching, pattern analysis, dynamic ranking, and a harder pass on queries that want experience, opinion and reviews. Snippets started using that analysis too. Trade press later reported a May 2021 consolidation of VIEW’s ranking modules into one model, with Naver saying C-Rank and D.I.A.+ criteria themselves were not being replaced. That report is not given a table row here. It is a plumbing change, and the primary post is less cited than the 2018 and 2020 essays.
Smart Blocks are how Naver then took VIEW apart. By 11 November 2022 the company was shipping FAQ and Q&A blocks: official answers for health, law and policy, and many-voiced answers for everyday questions, pulled from Knowledge iN, blogs and cafés, with AI snippets. The press note says FAQ blocks had already widened in June 2022. On 1 February 2024 Naver replaced the VIEW area in integrated search with Popular Posts Smart Blocks and split the old VIEW tab into Blog and Café. The press release is unambiguous: intent clusters, not a single blended VIEW box. C-Rank and D.I.A. did not vanish because the tab was renamed. They feed a different container.
The generative layer is recent and dated. AI Briefing launched on 27 March 2025 (Naver’s English release), summarising search and extending into shopping, places and short-form. AI Tab followed on 26 June 2026 for every main-page user (launch release). These are not secret ranking formulas. They are new surfaces that decide whether a blog is read, cited, or skipped. Early evidence on which queries earn a citation, and which commercial keywords still do not, is in the AI Briefing query note and the community recap.
Use the table as a map of public models, not as a promise that Naver stood still in the blank years. The blanks are 2006–2011, the stretch between SONAR and C-Rank, and dozens of collection-level tweaks that never earned a name. Same standard on Yandex and Baidu. Last reviewed 3 October 2026.