Poplink Ads

Overview

Built a contextual advertising system that analyzes blog/article content to display relevant ads without relying on third-party cookies or user tracking. A timely solution given the industry's shift away from cookie-based targeting.

How It Worked

Article Ingestion & Classification

Client articles were scraped and stored in our database, then processed through MonkeyLearn to generate context maps - classifying each article and extracting semantic meaning for individual words and phrases.

Ad Classification

Ads were similarly classified by context. This let us pre-compute which ads matched which articles based on contextual alignment, not user behavior.

ptag Integration

Publishers added a single script tag to their markup (similar to gtag). When loaded, ptag sent the page URL to our backend. We identified which article the user was viewing and determined which contextually-matched ads to serve.

Keyword Targeting

The script tracked which keywords were visible in the user's viewport. When a targetable keyword scrolled into view, we displayed the matching ad as a popup link on that term - hence "Poplink".

Click Funnel

When users clicked a popup link, they passed through our tracking funnel before reaching the advertised site. This recorded the ad click, attributed it to the publisher, and provided conversion data to advertisers - all without cookies.

Technical Highlights

• MonkeyLearn-powered article and ad classification
• Pre-computed context matching between articles and ads
• Lightweight embed script with minimal page load impact
• Real-time viewport tracking for keyword visibility
• Click attribution through redirect funnel
• Privacy-first: no user tracking or cookie storage