Tao Feng

0-to-1 Ads Product

Landing page destination optimization

An AI-assisted ads product that chooses the most useful post-click destination for a buyer at impression time.

The product bet

The same ad can lead to multiple destinations: product detail pages, shops, websites, apps, or other surfaces. The bet was that destination choice should become a personalized delivery decision instead of a static advertiser setup decision.

What made it hard

  • Advertisers needed simple controls, while the system needed enough flexibility to learn.
  • Launch sequencing mattered because eligibility, integrity, measurement, and GTM readiness moved at different speeds.
  • The product had to graduate from understandable heuristic logic toward model-based ranking without losing trust.
The lesson: when AI changes a decision, the PM still has to design the promise, the guardrails, and the path to trust.

Why it matters

Post-click optimization is a useful example of modern ads product work: the user experience, advertiser outcome, model objective, measurement plan, and sales narrative all need to converge before the product can scale.