How Machine Learning and Measurement Infrastructure Are Powering the Creator Economy at Scale — Featuring Lukia Yuanshu Xu

The creator economy—now valued at over $480 billion globally—runs on a layer of technical infrastructure that most people never see. Behind every brand partnership, every paid creator post, and every advertising campaign on a social platform sits a stack of brand-creator matching algorithms, ads delivery systems, and measurement infrastructure that determine whether the money flows, whether the results get reported, and whether advertisers come back. Building that infrastructure at scale, across two of the world's most influential social platforms, is the work that has defined the career of product leader Lukia Xu.

Xu is a Staff Product Manager at Pinterest and a former Group Product Manager at TikTok/ByteDance, where she spent six years building the foundational ads measurement and creator monetization systems that enabled TikTok to compete—and win—for enterprise brand budgets globally. Her work sits at the intersection of machine learning, large-scale experimentation, and product design: highly technical problems with direct revenue consequences at platform scale.

The Measurement Gap in Digital Advertising

For a platform to attract the largest brand advertisers—companies like P&G and Unilever—it must satisfy a checklist of technical requirements around viewability, brand safety, and measurement that has been standardized by industry bodies like the IAB and MRC. When TikTok began its international monetization push in 2018, it met almost none of them.

Xu, joining as TikTok's first product manager in ads measurement, identified that the IAB Tech Lab had recently introduced the Open Measurement SDK (OMSDK): a new standard that allowed publishers to conduct a single SDK integration that is able to support all compliant measurement vendors simultaneously, eliminating the need for individual server-to-server integrations with each vendor. The alternative—the approach being used by other platforms at the time—required separate, expensive integrations with each measurement company, and took months to deploy. Xu then proposed and led the OMSDK integration, working directly with IAB Tech Lab to define viewability tracking protocols for TikTok's unique full-screen, sound-on video format—a format that had no existing measurement precedent. ByteDance received IAB certification in November 2018, becoming one of the first Chinese companies to achieve it and the first major Chinese social platform to do so, unlocking enterprise brand budgets that had previously been inaccessible.

Designing a Native Survey System for Brand Ads Measurement

With viewability solved, the next gap was effectiveness measurement. Brand Lift Study (BLS) — the standard method for measuring whether an ad campaign changed consumer awareness, recall, or purchase intent—was available from third-party research firms, but at a cost of several thousand dollars per study and with significant operational overhead. For TikTok to scale brand advertising, it needed a first-party, automated solution.

Xu designed TikTok's Brand Lift Study system from the ground up, making deliberate technical choices at every layer. The survey delivery mechanism was a TikTok-native in-feed video experience with a custom pop-up survey card—leveraging the existing in-feed ad format. This approach allowed the survey to be served via the ads system, making collection control and cost measurement straightforward through existing infrastructure rather than building parallel systems.

Aside from the survey experience, the solution also included a reporting interface built natively into TikTok Ads Manager, allowing campaign managers to configure studies, select question templates, and view results without leaving the platform.

The system was deliberately designed to run at zero additional cost to advertisers, which Xu identified as critical and aligned with industry parity. The result: the number of campaigns TikTok was able to measure grew significantly from 2020 to 2021. Adweek covered the global rollout as a significant improvement to TikTok's measurement capabilities.

Democratizing Creator Marketing with Machine Learning and Automation

Branded Mission—described by TechCrunch at its launch as an industry-first creator marketing solution—presented a distinct class of technical problems. The product's core value proposition was enabling brands to run creator marketing campaigns at a scale that manual coordination could not support: thousands of creators per campaign, across dozens of markets, with automated workflow from selection through payment.

The matching system used machine learning to identify and invite relevant creators for each campaign based on a variety of signals. Once invited, creators submit videos directly through the TikTok camera following advertisers' prompts. Videos then first get distributed organically in the feed, which lets advertisers observe performance before selecting videos to be boosted as ads. Advertisers can view all submissions directly on ads manager, check video performance, creator profile, and boost with one button. The system then automatically turns the UGC into an ad and serves through the feed.

Under Xu's leadership, Branded Mission went from Beta testing to global GA, now available across more than 20 countries, and activating thousands of creators for each campaign.

Premium Ad Products at Pinterest

As Staff Product Manager at Pinterest, Xu now leads Premiere Spotlight, the company's flagship premium upper-funnel ad format. She owns end-to-end from strategy, format design, to advertiser experience, pricing, and reporting. As her journey continues at Pinterest, she is set to bring more impactful products by combining the highly intentful user journey with solid ads infrastructure.

Beyond her platform work, Xu mentors through ADPList and AIGA New York's Growth Programs, and was recently admitted as a Fellow of Beta Fellowship—a selective innovator community whose alumni include senior professionals from Google, Amazon, Meta, and TikTok, evaluated through an expert panel review on the basis of demonstrated field contributions.

"The interesting problems in ads infrastructure are never purely technical," Xu says. "Every architectural decision has a revenue consequence, and every revenue decision has an architectural constraint. The job is to hold both at the same time."

Lukia Yuanshu Xu is a Staff Product Manager at Pinterest. She holds a Master of Arts in Communication from the University of Southern California and a Bachelor of Arts in Economics from Peking University. Connect with her on LinkedIn: linkedin.com/in/lukiaxys/