- TikTok is building a solution that works like a data clean room called PrivacyGo.
- The tech is designed to smooth out advertiser-platform data sharing while guarding privacy.
- PrivacyGo may boost TikTok’s pitch to brands as regulations and privacy rules weaken ad targeting.
TikTok is developing a new privacy-tech solution for advertisers called PrivacyGo, a company spokesperson confirmed to Insider.
PrivacyGo will function similarly to a data clean room, where two parties with separate data sets like email addresses and sales information can match data in a privacy-safe way to find overlaps in their audiences. For example, a car manufacturer might compare its own prospective customer data to a segment of TikTok users who have watched car videos, target them with ads, and measure whether those ads drove results, like visits to a website.
Ad targeting and measurement have both become harder as privacy regulations and tech platforms like Apple have introduced more stringent privacy protocols, stripping away billions in ad revenue from companies like Snap and Meta Platforms. Clean rooms help drive ad performance and restore lost revenue. Tech giants like Google, Meta, and Amazon, as well as smaller adtech companies like Habu and InfoSum, offer clean-room tech to advertisers.
If TikTok’s approach to data anonymization makes it easier for brands to match audiences on its app, it could lead to a boost in ad revenue for the company.
Though TikTok is a much smaller ad seller compared to Google, Meta, and Amazon, it has been building solutions to entice advertiser budgets. In February, it rolled out new ad tools for small businesses — the types of companies that account for the most ad revenue on Meta and Google.
TikTok said PrivacyGo uses a type of cryptography called multi-party computation, or MPC. This tech, per the abstract of an upcoming presentation TikTok researchers plan to give on PrivacyGo, includes a new way of finding commonalities in multiple data sets without sharing the original sources. The abstract also mentions TikTok is using “differential privacy,” a technique where a data owner alters some inputs so that it can still be used for analysis, but can’t be exploited to identify information about individuals.
“The United States government uses it to allow people to analyze census data,” said Myles Younger, head of innovation and insights at the digital-marketing education firm U of Digital. “It was very important to have really strong cryptographic methods to allow people to work with that data while at the same time making sure that they couldn’t identify a bunch of individual American citizens and learn things about them that they really shouldn’t know.”
User privacy is a hot-button issue for TikTok, which has faced a flurry of accusations around how it stores and protects user data.
Government officials and former ByteDance staffers alike have raised concerns that the Beijing-based ByteDance could be forced to give the Chinese government access to US user data via its National Intelligence Law. TikTok has repeatedly said it has not given data to the Chinese government and would not if asked.
TikTok’s development of PrivacyGo may face similar scrutiny, though the company is making it available open-source and posting information about its techniques on GitHub, adding visibility to its data-matching process. The company has also said it plans to work with other open-source foundations down the road.
Read the full article here