Controversy is growing over a “world No. 22 for internet” ranking released by U.S. private firm Ookla. The government and telecom industry have issued a series of rebuttals after the survey found South Korea’s internet speeds were slower than those of Vietnam and Thailand. Critics say the measurement conditions and detailed data underpinning the rankings have not been sufficiently disclosed, limiting their use as a benchmark for assessing national network standards.

According to the telecom industry on Oct. 19, South Korea’s fixed ultrafast broadband download speed was 251.33 Mbps, ranking 22nd in the world in Ookla’s August Speedtest Global Index. Vietnam ranked 8th at 301.48 Mbps and Thailand 9th at 298.92 Mbps, placing both above South Korea.

The findings have fuelled controversy over insufficient investment in domestic telecom networks. Questions have also been raised over whether the measurement method and the representativeness of the data, which form the basis for the rankings, have been sufficiently verified.

◆ Ookla measures global internet speeds; OECD also points to limits

Ookla is an internet speed measurement and analytics firm founded in the United States in 2006. It operates Speedtest, which allows users to check download and upload speeds and latency. Based on user measurement data collected worldwide, it publishes monthly country-by-country internet speed rankings.

Its influence is significant enough that the International Telecommunication Union (ITU) and the Organisation for Economic Co-operation and Development (OECD) use related data to analyse telecom services. But that does not mean Ookla’s country rankings are officially certified.

A limitation of Ookla is that it relies on users voluntarily running speed tests. Because it gathers data from users who choose to participate, rather than designing a sample to represent subscribers nationwide, there is a risk of sample bias.

In its July report last year, “Closing Broadband Connectivity Divides for All,” the OECD pointed to the possibility of selection bias in user-driven measurements. It said users with internet connection problems or users more familiar with technology could take part in measurements in relatively large numbers.

In that report, the OECD excluded South Korea from a chart comparing regional fixed internet download speeds by country for the fourth quarter of 2024 using Ookla data. It cited the number of measurement samples. That does not mean Ookla’s overall country rankings were wrong, but it shows there are constraints on how the data can be used.

◆ “Just changing servers can mean a two- to three-fold difference”; test conditions are opaque

Another issue is that results can vary depending on the environment, including the test server and the subscription plan. A Ministry of Science and ICT official explained that tests specifying a server in South Korea and a server in Tokyo produced speed gaps of as much as two to three times in some cases.

The official added, “We cannot know how it controls measurement conditions,” and stressed, “We cannot say Ookla is 100 percent wrong, but it is questionable whether it is appropriate to accept opaque measurement-condition data as it is and evaluate country-by-country rankings.”

Even so, it is not easy for the government to verify broadband speed rankings across countries on its own. The ministry conducts annual evaluations of telecom service quality in South Korea, but the scope is limited to the domestic market. Measuring internet networks worldwide under identical conditions would require local survey cooperation, separate measurement standards and a large budget.

The official said, “To conduct quality evaluations on a global scale, multiple conditions are needed, including methodology and budget,” adding, “It is not realistically an easy issue.” That is also why the government, while disputing Ookla’s findings, cites OECD metrics such as fibre-optic penetration and the share of high-speed internet subscribers.

The telecom industry says detailed data used in measurements must be disclosed first to judge the credibility of the rankings. The Korea Telecommunications Operators Association (KTOA) said it was difficult to check detailed materials such as the number of measurements by country, the composition of subscription plans and the distribution of measured speeds, creating clear limits in comparing ultrafast internet quality between countries. It said it asked Ookla for detailed survey materials in 2023 but did not receive specific data.

KTOA plans to again demand that Ookla disclose its measurement methods and conditions and detailed country-by-country data.