Naver has changed how it ranks restaurants, and the new signals are what diners do, not what restaurants write. Since 1 October, local restaurant searches in parts of Korea show two new sections on Naver Place: themed recommendations (테마별 추천) and restaurants with fast-rising interest (인기 급상승 맛집). These sit alongside the familiar list. Both are driven by context and behaviour rather than by anything a restaurant can type into a profile field.
What changed
According to NetContentSEO’s write-up of the SmartPlace notice, Naver published the announcement on 2 October and the rollout began on 1 October, after A/B testing. i-boss reported that test back in March. It ran on mobile in selected areas from the end of March and used location, menu, time of day, weather and visit purpose together with directions, reservation and review data.
The production version keeps that recipe:
- Themed recommendations reflect user preferences and context: area, menu, time of day, weather and purpose of visit.
- Rising-interest restaurants use behavioural signals: directions requests, reservations, reviews and saves.
- Coverage is limited for now. The launch areas listed are Gangneung, Gyeongju, Gwanggyo, Gwanghwamun, Daejeon, Bundang, Seongsu, Yeosu, Jamsil and Haeundae.
i-boss’s 6 October round-up gives the practitioner reading. Restaurant visibility is expanding into personalised recommendation, so managing real user reactions (reviews, saves, bookings) matters more than ever.
Naver hasn’t published signal weights or a formula. It also hasn’t called this a generative AI feature. It is a recommendation layer on top of Place, not an AI Briefing for food.
Local SEO in Korea just became more behavioural
In Google-land, local SEO still leans heavily on the business profile: categories, proximity, reviews, citations. Naver is making explicit that the next layer of restaurant discovery runs on what people do with a listing: route to it, book it, save it, review it. And context like a rainy evening or a lunch-hour search decides which theme you appear in.
That rhymes with the rest of Naver’s product direction. AI Tab hands searchers off to maps and bookings. i-boss notes that AI Tab passed 10 million monthly users in early August, with commerce and local click-through rates above 40%. Place is where intent turns into a visit, and Naver wants to rank that moment with its own first-party data.
The obvious problem: behaviour can be bought
The awkward timing is that Naver is in a public fight over exactly this kind of signal. As Kukmin Ilbo reported on 5 October, Naver asked police in February to investigate Toss’s rewarded “visit mission” marketing. Naver argues that searches people are paid to make inflate search volume and can distort rankings on Place and Store. Any system that rewards saves, routes and reservations will attract people trying to manufacture them. Naver’s anti-abuse team now has a much bigger surface to defend.
For restaurant owners and agencies, that cuts both ways. Fake saves and incentivised review drives may move the rising-interest module briefly. They are also exactly what Naver has just told the police it considers abuse.
What to actually do
- Get the basics exact. Menu, opening hours, booking options and visit details should describe the real experience, because the themes match on context like time and purpose.
- Make booking and directions effortless. If reservations and route requests are signals, friction on Naver Reservation or a wrong map pin costs you twice.
- Earn saves and reviews honestly. Encourage genuine customers to save and review, and leave the rewarded “search mission” schemes alone.
- Monitor the new sections separately. In the ten launch areas, record the classic list, each theme and the rising section with dates and screenshots. A single ranking position no longer tells the story.
Our view: this is Naver doing what Naver does best, using data nobody else has. It is good for diners and probably good for genuinely busy restaurants. It is bad for anyone whose local SEO plan was a keyword-stuffed Place description. Expect the themes to spread beyond the first ten areas, and expect the first wave of “rising restaurant” manipulation offers on Korean marketing forums soon after.