This post was written by Yaniv Avraham, Principal Data Product Manager at Start.io.

Start.io‘s audience portfolio already helps advertisers reach users by demographics, interests, intent and more. Now we’re adding a powerful new family of segments to our taxonomy: Performance Audiences – dynamic, behavior-based segments built entirely from first-party ad interactions across Start.io‘s massive supply network. 

Instead of inferring who a user might be, Performance Audiences capture what users actually do when they encounter ads – how they click, how much demand they attract, and how they engage across formats. 

Why This Is Different 

Many DSPs already let buyers optimize campaigns toward performance KPIs like click-through rate (CTR) or bid rate. But that optimization typically happens behind the scenes, inside the DSP’s own algorithm – and it’s limited to the traffic that single DSP sees. 

Start.io‘s Performance Audiences are fundamentally different. We observe user behavior across our entire exchange network, spanning massive integrations with SSPs and DSPs alike. That means the audiences we deliver aren’t shaped by one bidder’s narrow view of the market. They’re broader in coverage, less biased, and more holistic – reflecting how users behave across the open programmatic ecosystem, not just within a single demand-side silo. 

What’s Available at Launch 

Performance Audiences launches with two core dimensions, each offering tiered segments: 

Click Performance ranks users by real click-through behavior over a rolling window: 

  • High CTR Users – the top 5% of clickers across display and video. First-party validated engagers, ideal for direct-response campaigns optimizing for clicks and CPA. 
  • Above-Average CTR Users – a broader pool of consistently engaged users, giving performance buyers scale without sacrificing engagement quality. 
  • Ads-Exposed Non-Clickers – users regularly served ads who don’t click. A powerful suppression layer for throttling optimization. 

Bid Value surfaces the users the programmatic market itself values most: 

  • High-Bid Users – users attracting top-tier winning bids across Start.io supply. A strong proxy for commercial intent and competitive demand. 
  • Above-Average Bid Users – a larger reach pool of users consistently drawing premium bids, enabling advertisers to tap into high-value demand at scale. 

Built for Precision 

Every segment is computed per country and OS, so thresholds reflect local market dynamics – a “High CTR” user in the US is benchmarked against other US users, not averaged globally. Segments refresh automatically as behavior changes, keeping audiences fresh and actionable. 

What’s Next 

This is Phase 1. In upcoming releases, we’ll expand Performance Audiences in two directions: 

  • Ad format granularity – each performance segment will be available per ad format, so buyers can target, for example, High CTR users of video separately from High CTR users of display banners. This lets advertisers match audience selection to their specific creative strategy and campaign format. 
  • Ad category targeting – performance segments broken down by ad vertical, enabling advertisers to reach users based on their engagement with specific ad categories. For example, target users who clicked on automotive ads, or those who engage heavily with finance or CPG creatives. 
  • New performance dimensions – including video completion rates, viewability, conversion signals, and user freshness – giving buyers even more ways to optimize toward real outcomes using Start.io first-party data. 

Performance Audiences – when Ad-performance meets audience data.