Home > Publications > Digital technologies & governance > When geo-targeting technology helps:  A brief case study on ‘Douyin Missing People’

When geo-targeting technology helps:  A brief case study on ‘Douyin Missing People’

The role of social media in finding missing people
Jufang Wang

Jufang Wang

Deputy Director of Oxford Global Society and the coordinator of its Digital technology cluster

Share this post

Share on facebook
Share on twitter
Share on linkedin
Share on email

The views expressed are solely those of the author (s) and not of Oxford Global Society.

We have been increasingly warned of the dark side of personalised or targeted information, such as ‘filter bubble’ and ‘surveillance capitalism’. However, this technology has great potential in delivering social good.  On 15 February 2023, ByteDance, the parent company of TikTok, released a video in which its CEO Rubo Liang announced that the company’s platforms have helped to find over 20, 000 missing persons in China since 2016.

ByteDance’s achievement in locating missing persons brings up some questions. Why has only ByteDance announced such a level of success in finding missing persons? Why hasn’t this happened to Facebook, Twitter and YouTube? What are the factors behind ByteDance’s success in this public endeavour? Can ByteDance’s model be adopted by others?

The technology is not new

ByteDance launched its Missing People programme in February 2016. Initially called ‘Toutiao Missing People’, the programme was later renamed ‘Douyin Missing People’, as Douyin (Chinese version of TikTok) surpassed Toutiao (an algorithm-based news aggregating platform) to become the company’s most popular platform. Soon after its launch, the programme established a partnership with the Department of Social Affairs (DSA) of the Ministry of Civil Affairs of China, which has around 2,000 local institutions dealing with reports of missing persons. Apart from cases from the DSA, ByteDance’s project team also actively responds to missing person cases reported by individuals. According to ByteDance, in the last 7 years, it has released over 170,000 pieces of information about missing persons and helped to find over 20,000 persons in China (around 12% successful rate).

So, what is ByteDance’s mechanism in finding missing persons? The technology used is quite straightforward: geo-targeted information. That is, using its popular platforms, including Douyin and Toutiao, ByteDance pushes the information about the missing via pop-up windows to users believed to be at the same locations.

Geo-targeted information is not new for government departments or social media platforms. For instance, during or before anticipated natural disasters, it is common for governments to send out geo-targeted warnings to relevant residents, with the help of telecommunication operators. Social media platforms like Facebook have also partnered with national governments to push geo-targeted alerts in some serious cases, such as those concerned abducted children.  

Factors that mattered

If the technology used by ByteDance is nothing new or innovative, what are the main reasons that have led to its astonishing results in finding missing persons?

One important factor may be ByteDance’s high level of involvement in, and commitment to, its Missing People programme. As Mr. Liang, ByteDance CEO, emphasised in the video mentioned, the company had committed itself to this Corporate Social Responsibility (CSR) programme as it would to any other profitable ‘products’. In addition to providing a platform that users can post information about missing persons, the company itself plays an active role by pushing information to targeted users. Given the immense reach of Douyin and Toutiao in China, it is not surprising that the company has achieved great results over the years. As of August 2020, Douyin had over 600 million Daily Active Users (DAU) in China.

In comparison, while major US-based social media platforms such as Facebook and Twitter have been important tools for spreading the information about missing persons, they mainly play an intermediary role. That is, users, including police departments, dedicated civil organisations, and individuals, can spread the word among their friends or followers, but platforms themselves usually do not actively recommend or push such information. These platforms do have partnerships with government departments or civil organisations to actively push missing people alerts in their news feeds, but this happens only in serious cases rather than each case.

Douyin’s short-video format and its unique design features may be another main factor that has contributed to its effectiveness in helping to locate the missing. Earlier, staff members of the American National Center for Missing & Exploited Children found that video posts were more effective than those with a photo and written description of the missing. Additionally, on Douyin, users have to scroll videos and cannot just skip a video recommended by the platform, which means missing person content can have better chance of being seen. An earlier study found that missing person videos can have a very high level of engagement on TikTok (e.g., one video had 2m views, 1.3m likes, and 13.5 comments), while most missing person posts on Twitter only have around 100 views. Douyin and TikTok are almost identical in many aspects including its user interface designs, although Douyin is available only in China mainland and TikTok only in overseas markets.

Lastly, to push geo-targeted information, platforms must gather location data from their users. Both Douyin and Toutiao offer a feature called ‘Tong Cheng (or living in the same city)’, which recommends posts related to users’ locations. To utilise this feature, users have to grant the platforms permission to collect their location information. This design and the relatively loose policy environment regarding personal information in China may have been a factor behind ByteDance’s successful story in finding missing persons.

Possible wider application

It has been reported that in certain countries, a child is reported missing almost every minute. Since ByteDance has been successful in helping to locate missing persons at a relatively high rate, it is worth considering whether its model can be replicated elsewhere.

One lesson from the ByteDance case is that platforms’ active involvement in missing person cases can be crucial. With hundreds of millions, or even billions, of users, an enormous amount of collected data, and advanced proprietary technologies, major social media platforms possess an unparalleled capability in spreading information to a wider and more targeted public. For this reason, if platforms like Facebook and Twitter play a more active role in pushing geo-targeted information about missing persons (like ByteDance), positive results could be expected.

However, ByteDance’s model in pushing missing persons information may be deemed as a bit intrusive (as it uses pop-up windows) by many users. Additionally, location data is a type of sensitive personal information, and many users may not be willing to share it with platforms, making precise geo-targeting impossible. These may be the reasons why TikTok has not yet committed to the Missing People programme like its sister platform Douyin. However, there are possible solutions to overcome these obstacles. In fact, to push geo-targeting information about missing persons, platforms do not need to collect the precise location data of users, and rough location information is sufficient. As a result, platforms may reach agreement with users in advance that their rough location information can be used to push missing persons alerts.

As for other concerned parties, such as governments, one option is to partner with major platforms in finding missing persons. It may be worthwhile to explore whether such partnerships can go beyond serious cases (e.g., abducted children) and extend to wider cases, such as missing elderly persons with dementia. For government departments, another choice is to use geo-targeted text messages to reach the public about the missing. A case in point is how the NSW police in Australia have been using this method to solve missing person cases and has so far achieved great results. It is reported that, after using this method, the state had only 18 missing person cases reaching 90 days and obtaining long-term status in 2020, compared to an average of 147 cases reaching that status between 2015 and 2019.

It is worth noting that this article is only a brief and initial study of the ByteDance Missing People programme. More systematic and rigorous research on this issue should be able to deliver more insights, from which more useful and evidence-based policy recommendations may be drawn to inform governments and other concerned parties.