Google Overtakes Naver in Korean App Users: How the Numbers Were Built
In July 2026 the ranking of Google and Naver by monthly active users (MAU) in South Korea was reversed for the first time. The figure comes from Mobile Index, a data service operated by IGAWorks, and was released on 11 August 2026. The headline is a simple reversal, but understanding how the number is produced changes how it should be read.
MAU is not a government statistic. It is an estimate produced by a private analytics firm from a sample, and each firm applies its own panel and its own definitions. The clearest evidence is that ChatGPT's user count for the same month differed by more than seven million between two research firms.
This article concerns the method behind the figure rather than the size of the figure itself.
Table of Contents
- What Monthly Active Users Actually Counts
- The Scope of the Mobile Index Estimate
- The July Picture in Data
- Why Research Firms Report Different Figures
- How AI Search Changed the Entry Point
- What This Indicator Cannot Tell Us
- Frequently Asked Questions
- Summary
1. What Monthly Active Users Actually Counts
Monthly active users, or MAU, refers to the number of people who opened a given application at least once during a calendar month. Three things are excluded from that definition.
Frequency is excluded. Someone who opened the app once in a month and someone who opened it ten times a day both count as one. Duration is excluded. A session closed after three seconds is not distinguished from thirty minutes of use. Channel is also excluded. Visits made through a web browser rather than the application are not counted in this figure.
A comparison of Google and Naver by MAU therefore answers the question of which application was opened at least once on a larger number of smartphones. It does not answer the question of which service was used more.
2. The Scope of the Mobile Index Estimate
The figure was published by Mobile Index, a service operated by IGAWorks (a South Korean mobile data and marketing analytics company). It collects application usage data from Android and iOS smartphones in South Korea and produces a statistical estimate of the total user base. The panel design and the adjustment algorithm have not been disclosed in detail.
The series begins in March 2021. The word "first" refers to the period since that date; the situation before it cannot be verified from this source.
The classification also requires attention. The first and second positions cited here are ranks within the productivity app category, not across all applications. Across all applications, YouTube and KakaoTalk rank ahead, with Google third and Naver fourth. The two exchanged places against the previous month, each moving by one position.
3. The July Picture in Data
Monthly active users in the productivity app category, July 2026
| Application | Users | Category rank | Note |
|---|---|---|---|
| 47,022,854 | 1 | First time at the top of the series | |
| Naver | 46,845,017 | 2 | Ranked first in the previous month |
| Google Chrome | 42,457,506 | 3 | Web browser |
| ChatGPT | 16,456,151 | 7 | Passed Daum in April 2025 |
| Daum | 6,307,395 | — | About 38 per cent of ChatGPT |
Source: IGAWorks Mobile Index, data as of July 2026. The gap between Google and Naver is 177,837 users, equivalent to roughly 0.4 per cent of Naver's user base.
The more striking figure is not the gap at the top but the drop between the fourth and fifth rows. ChatGPT stands at approximately 2.6 times the size of Daum. The migration of portal users towards conversational AI appears on a far larger scale than the reversal at the top of the table.
Composition of AI chatbot app users, July 2026
Source: WiseApp Retail, data as of July 2026
In the WiseApp Retail count, ChatGPT led with 23.67 million users, followed by Google Gemini at 9.47 million. Zeta recorded 4.28 million, Claude 3.44 million, Perplexity 1.47 million, Grok AI 1.37 million and A. 1.15 million. Gemini, Zeta and Claude each posted their highest monthly user counts since launch.
4. Why Research Firms Report Different Figures
At this point the same application in the same month produces two different numbers. ChatGPT's July 2026 user count is 16,456,151 according to Mobile Index and 23.67 million according to WiseApp Retail, a difference of roughly 7.21 million.
| Item | Mobile Index | WiseApp Retail |
|---|---|---|
| Operator | IGAWorks | WiseApp Retail |
| July ChatGPT figure | 16,456,151 | 23.67 million |
| Panel size | Not disclosed in detail | 36.61 million Android, 14.61 million iOS |
| Scope | Productivity app category ranking | AI chatbot apps with at least 200,000 monthly users |
| Exclusions | Web usage | Web usage; apps centred on call or document summarisation and voice recording |
Source: materials published by each firm and reports citing them, data as of July 2026
The divergence reflects design differences rather than error. Each firm draws its panel differently, includes different age groups and devices, and applies its own weighting to produce national estimates. Because neither firm publishes its full methodology, the precise cause of the gap cannot be identified.
The practical rule is straightforward. Compare a time series only within a single research firm, and do not calculate rates of change across firms.
5. How AI Search Changed the Entry Point
The change on Google's side coincided with a redesign of the search interface itself. In January 2026 Google switched the default model behind AI Overviews, its search summary feature, to Gemini 3, and in May 2026 it upgraded the default model behind AI Mode, its conversational search service, to Gemini 3.5 Flash. According to Google, AI Mode has passed one billion monthly users worldwide. A separate figure for South Korea has not been disclosed.
Naver has moved in the same direction. It has operated AI Briefing, which summarises search results, and on 26 June 2026 it formally launched AI Tab, a conversational AI search service, placing it in the main search field on both mobile and desktop.
On the revenue side the results are documented. Naver reported consolidated second-quarter 2026 revenue of KRW 3.3888 trillion (approximately USD 2.40 billion at the exchange rate of 12 August 2026, KRW 1,412.6 per USD), a record quarterly figure and an increase of 16.2 per cent year on year. Operating profit fell 0.2 per cent to KRW 520.3 billion (approximately USD 368 million at the same rate), which the company attributed to AI infrastructure investment. Naver stated that more than 60 per cent of its advertising revenue growth came from AI applications, and that AI Briefing advertising, formally launched in late July, recorded a click conversion rate more than 30 per cent higher and a purchase conversion rate more than three times higher than conventional search advertising. These figures are company disclosures and no third-party verification has been published.
The user-count indicator and the revenue indicator are therefore pointing in different directions. Neither alone is sufficient to judge the competitive position.
6. What This Indicator Cannot Tell Us
First, search market share. MAU counts people who opened an application, not search queries. This data cannot support a claim that search share has been reversed.
Second, intensity of use. In the WiseApp Retail data, Zeta recorded 4.28 million monthly users, fewer than one fifth of ChatGPT, yet its total app launches in July reached 2.008 billion against ChatGPT's 1.817 billion. Rankings by user count and by intensity of use do not align.
Third, symmetry of the units being compared. Electronic Times (etnews), citing an industry source, reported that because Naver also operates separate applications such as Naver Blog, Naver Cafe and Chzzk, placing a single Naver app beside a single Google app has limits as a comparison. App-level rankings depend on how widely a company distributes its services across separate applications.
Fourth, persistence. The gap of 177,837 users amounts to about 0.4 per cent of the total, a margin that a sample-based estimate could reverse the following month. The Korea Economic Daily, citing an industry source, reported that it is too early to treat one month of app user counts as a reversal of leadership in the search market.
7. Frequently Asked Questions
Q. Does this reversal mean fewer people are using Naver search?
A. No. Naver's app MAU has moved within a range of roughly 43 million to 46 million from January 2023 through July 2026. The figure has been flat rather than falling, and Google rose into that range. Google's app user count passed 40 million in May 2025 and 46 million in May 2026.
Q. If research firms disagree, which figure should be used?
A. Neither is definitive. Both are sample-based estimates built on different designs. The safer approach is to select one firm appropriate to the purpose and track only that firm's time series.
Q. Would the ranking change if web usage were included?
A. This cannot be determined. The published data covers mobile applications only, and no combined indicator including web usage has been released.
8. Summary
In July 2026 Google recorded 47,022,854 monthly active users in South Korea against Naver's 46,845,017. It is the first such ordering since Mobile Index began the series in March 2021.
What matters is not the ranking for a single month but an understanding of the measurement structure behind it. If the definition of the indicator, the identity of the research firm and the scope of exclusions are not verified, the same source can support conflicting conclusions. Before citing a number, establish what that number did not count.
Investment Disclaimer
This article is provided for information and analysis based on publicly available materials, and does not recommend the purchase, sale or holding of any financial product or real estate asset.
The figures and outlooks cited here reflect the sources available at the time of writing and may change thereafter. Regulations, tax rates, interest rates and eligibility rules apply differently depending on the effective date and on individual circumstances.
All investment decisions and any resulting gains or losses are the sole responsibility of the investor. Please consult a qualified financial, tax or real estate professional before acting on this information.
댓글
댓글 쓰기