Understanding TikTok Data Privacy and Trust User Interviews

Background

Team: Trust and safety

role: end to end lead researcher

Timeline: 6-7 weeks

product: TikTok trust intiatives

This was a strategic foundational project that aims to understand TikTok users’ trust journey and identify areas of opportunity for the product and marketing teams.

Methodology: moderated user interviews

While TikTok had Trust Benchmark Surveys and Brand Health Tracker Surveys, there was no qualitative understanding of what the survey numbers mean. These interviews provided color and life back into the quantitative survey results.

Objective

One of the top goals for TikTok leadership is to become one of the most trusted social media platforms. This initiated a Trust Benchmark and Brand Health Tracker Survey that consistently measured trust across different regions (e.g., US, UK, MX, JP, FR, DE, and ID). Insights from the benchmark showed that some of the top drivers of trust are users’ perceptions of data privacy policies, however users were scoring in the mid-low ranges. This study aims to understand what shapes users' trust, and how we can remediate mid-low levels of trust by understanding how we can improve perceptions of data privacy policies. 

research questions

1. How can we remediate mid-low levels of trust, if anything?

2. What shapes users trust in TikTok?

3. What are key drivers of trust across key regional markets?

Methodology

Participants

• Monthly Active TikTok Users

• Half the users are self reported high trust in TikTok

• Half the users are self reported low trust in TikTok

• Internal Portion: I internally conducted user interviews with a total of (30) TikTok users in the local languages

  • 10 users from United States (English)

  • 10 users from United Kingdom (English)

  • 10 users from Mexico (Spanish with support from another researcher)

• External Portion: I managed an external vender to conduct user interviews with a total of (50) TikTok users in the local languages

  • 10 users from United States (English)

  • 10 users from Japan (Japanese)

  • 10 users from France (French)

  • 10 users from Germany (German)

  • 10 users from Indonesia (Indonesian)

rational

I chose to recruit high trust users to understand what builds trust and how they overcame trust barriers if any. 

I chose to recruit low trust users to identify moments where trust breaks down and uncover opportunities to remediate low levels of trust.

I chose to use a vendor because they provided broader access to international participant samples and operational support (e.g., recruiting, moderation in local languages, translations, etc). 

US was included in both internal and external because the topic was more emotionally charged and nuanced given the potential TikTok ban in the US. This allowed insights be more objective, without interactions with internal employees, and doubled the sample size allowing for richer insights.

Procedure

• Introduction (5min): Provide instructions and built rapport

• General Trust in Apps (10min): Understand their general trust in apps, specifically social media apps. Dive into any positive or negative experience they may have had with data privacy policies from any app.

• Trust in TikTok (15min): Understand their overall trust journey with TikTok and any concerns they have with TikTok’s data privacy policies.

• Data Initiatives (15min): Reviewed TikTok data privacy initiatives to understand users’ perceptions and whether education of the different initiatives impact their trust in TikTok.

• Data Privacy Features (10min): Understand their perceptions of TikTok’s data privacy product features and how they compare to other platforms.

• Wrap-up Questions (5min): Follow-up questions were asked to allow participant to reflect on their overall experience and identify if education of data initiatives or features changed their trust in TikTok.

Application: Zoom interviews

• Internal interviews were recruited via backend query pulls, which sent an in-app notification to their TikTok account, inviting them to participate in an interview for a gift card.

• External interviews were recruited with support from the vendor.

Key Findings

1. trust is easily lost and hard to gain

• Typically users start off with moderate trust. They may have heard negative news in the media about TikTok, but they are curious about the app enough to give it a try.

• Users develop high trust in TikTok with more time spent on the app. High trust users understand that TikTok may be using their data to provide a more personalized feed, but see no issues with this equal exchange of data. This can take many months-years before a user reflects on their overall experience on TikTok and deems it overall positive.

• Users can easily lose trust in TikTok when they see bad actors (e.g., scams, fake profiles, inappropriate content, etc). These bad actors can reinforce what they previously have heard about TikTok and cause users to lose their trust quickly.

• However, TikTok has the opportunity to remediate mid-low levels of trust. Taking action on bad actors (e.g., removing fake profiles, flagging scammers, etc) can push low trust users to moderate trust users. And educating users on the many data privacy initiatives that TikTok has in place can also push a moderate trust user to a high trust user.

• Across all segments, most participants wanted TikTok to help them feel informed and in control of their data, but didn’t actively seek out information on trust or data privacy.

Key Recommendations

🛡️Prioritize Resolving Trust Incidents

1. Consider proactively responding to trust incidents on TikTok (e.g., reported scams, fake profiles, misinformation, etc).

2. Consider maintaining transparency by keeping the reporting user in the loop as the investigation progresses.

🔐Enhance Data Privacy Awareness & Control

1. Consider improving education of data privacy policies (i.e., Project Texas, Oracle Data Storage, etc) by leveraging creators, posting internal videos, marketing trust and safety website links, etc.

2. Consider ways to help users feel more in control of their data (e.g., consent to changes in terms and conditions, education on two factor authentication, etc)

2. First and second hand experiences can impact trust

• First hand personal experiences impact high and low trust users. High trust users anchor on their overall positive experiences on TikTok, while low trust users focus on the few negative experiences they have personally seen on TikTok.

• Second hand experiences impact moderate and low trust users. Moderate trust users are skeptical of all social media apps, and so their capacity to trust any social media app is overall lower, regardless if they have had only positive experiences on TikTok. Negative personal experiences of TikTok from low trust users are reinforced by what they hear from the news and media of all social media companies and the overall geopolitical climate.

• Overall, most participants felt TikTok was unknown and impersonal. Other apps felt more familiar because they could put a face to a name (e.g., Zuckerberg was associated with Facebook, Tesla was associated with Musk, etc.)

Key Recommendations

📢 Own the Narrative:

1. Consider proactively addressing negative news from media and press through channels such as marketing campaigns and TikTok content on the FYP

2. Consider humanizing the brand by associating a spokesperson with TikTok, helping users associate trust and accountability with leadership

3. Regional trust is driven by openness to foreign ownership, confidence in gov protections, and cultural compatibility

Key Recommendations

📝 Tailor Regional Messaging: Consider developing region-specific trust strategies so messaging and marketing can feel more contextual with cultural nuance (e.g. US can be more tailored to data transparency vs EU/UK might emphasize regulatory compliance and government alignment).

Key Recommendations

Impact