Korean marketers’ forum questions have moved since our last Naver community recap on 6 October. That one was about the “they patched Latest again” complaints and Naver’s fight with Toss over search abuse. In the nine days since, the threads on i-boss’s Q&A board have shifted to new subjects: Powerlink ad positions that depend on who is looking, organic visibility that disappears without any warning in Search Advisor, Clips taking keyword slots that blogs used to hold, and a blog-account trade that now includes year-long rentals. None of it has an official Naver explanation yet. Here is what people are reporting and how much of it holds up.
“My Powerlink rank depends on who is logged in”
Two separate threads describe the same thing. On 7 October an advertiser asked why the same keyword gives a different Powerlink position when logged in and when logged out, and a different one again for each Naver ID they test with. On 12 October another posted numbers. Logged in, the ad ranked 25th for users in their 30s and 40s, 6th for users in their 20s, and didn’t appear at all for users in their 70s. In incognito it came 3rd, and in a normal logged-out browser 6th. The poster suspected competitors, such as course and academy advertisers, were setting age-based bid weights. Their team is now discussing splitting campaigns by age group.
That guess is probably right. Naver search ads let advertisers apply gender and age bid weights at ad group level. Agency GI Corporation’s walkthrough on OpenAds gives the maths. A 10,000 won bid with a 150% weight for men becomes a 15,000 won bid when a man searches. Weights multiply, so 10,000 × 150% × 50% for 30-somethings gives 7,500 won. Users who aren’t logged in, or haven’t consented to data use, count as “unknown”. Your “real” rank is therefore a different auction for every demographic. A single logged-in check from the office tells you where you rank for one 34-year-old, not for the market.
There is also a ceiling. The same guide notes Naver bids are capped at 100,000 won even after weights. An older thread from an advertiser running law-firm and clinic campaigns complains that 100,000 won still doesn’t buy first place on some keywords. When everyone in a vertical sits at the cap, quality index and creative decide the order. Bidding can’t.
The same confusion is spreading to other ad products. On 13 October a marketer running Naver Brand Content ads in the interior DIY category asked why raising the bid by more than 1,000 won still didn’t get them into the top three. Was a rival simply bidding more, or does the quality of the copy count? On 14 October a Place search ad buyer asked how those ads are ranked at all, saying they don’t show even with bid weights at the maximum. The answer to both is the same. Naver’s ad auctions combine bid, weights and a relevance or quality score the advertiser can’t see. If you spend more without getting higher, the score is probably what’s holding you back.
“Three million indexed pages, and site: shows nothing”
The most alarming post of the week came on 12 October. A long-running site owner says Naver’s site: search first dropped to five or six results, then to none. Search Advisor still reports roughly 3 million indexed pages, and new posts used to appear within 30 minutes. The owner says robots.txt and Cloudflare are fine, and that Naver’s Yeti crawler is still visiting. Pages are being fetched but not shown.
We can’t diagnose a site we haven’t seen. The pattern, though, is crawl working, index count unchanged and visibility at zero, and in Naver that usually points to a site-level visibility restriction rather than a technical indexing fault. The site gets suppressed after a spam or low-quality review, and the Search Advisor index counter doesn’t notice. Naver doesn’t always send a message when that happens. If you see the same thing: check Search Advisor for any notices, run a handful of exact-title searches, stop changing the site, and file a re-review request with dates. Rewriting thousands of pages the morning after a drop just wipes out the evidence you need.
Blog tab caps, Clips and dormant blogs
- A three-post cap in the Blog tab. An interior-design marketer on 11 October blames what they call a three-post exposure limit in the Blog tab for a big hit. Smart Blocks are “tied up”, positions swing from first to far down overnight, and weak posts rank above good ones. The question is whether opening a second blog in the same category would count as abuse. Almost certainly, if it exists to get round a per-source limit.
- Clips taking keyword slots. A 13 October thread notes that Naver Clips now show for “many keywords”. Some are sorted by recency. Others are old, low-view clips that still rank first. That fits the trend in our AI Briefing citations recap, where Clips were among Naver’s own surfaces gaining share. Short-form video is now a separate ranking game inside integrated search, and nobody outside Naver knows its rules.
- The “72-hour” delay for dormant blogs. A user asks whether a blog with no new posts for three months still waits about 72 hours before a fresh post shows. That is folklore, not documentation. Naver’s help pages say up to 24 hours for new or edited posts. It is still worth logging publish times if you take over a quiet client blog.
- Ads that stop serving after a day or two. On 12 October a search advertiser on auto-bidding reports creatives that serve for a day or two and then stop. They ask whether sales history or bid level is to blame. With demographic weights and Naver’s growing ad-product mix, including the Trend Pick move to ADVoost, there are more moving parts than one bid.
The black market is still trading
Wednesday’s strangest thread: someone shopping for a blog account asks why one seller wants 50,000 won for a “준최4” blog and another 500,000 won for a “준최5”. Both are non-real-name accounts. The “준최” grades are unofficial semi-optimised tiers that marketers use to price accounts. Naver doesn’t publish any such scale, and its terms don’t allow accounts to be transferred. A bought blog that suddenly starts posting about interior-design reviews is exactly what the abuse filters are built to catch. That may explain some of the “posts vanished” complaints in the threads above.
The rental market is the other side of the same trade. On Thursday morning a blog owner described renting out an unused “optimised” blog after repeated KakaoTalk offers. The deal was 2 million won for a year, which turned into roughly 13 months once a “test period” was added. A third-party grade checker still rates the blog “optimal+1”. AI chatbots told the owner such blogs go for 3–7 million won, and they now want to rent it out again. The renter gets a blog that has history and trust, while the owner keeps the account and most of the risk. That is exactly the abuse pattern Naver’s spam policies describe. When the renter’s posts get the blog downgraded, the owner is the one with a dead asset.
Opinion: this week’s threads come down to one point. Naver results are now personalised and pieced together from many surfaces, so one person’s view of the SERP isn’t a measurement any more. For ads, check positions by demographic, or trust the reports over screenshots. For organic, keep timestamped logs before you blame an algorithm. And for anyone tempted by the account market, whether buying or renting, the cheapest blog is the one Naver hasn’t already flagged. Neither 500,000 won nor 2 million won a year buys you any guarantee of that.