Hyper-local performance at enterprise scale doesn’t just require better data, it requires execution models that can actually support it.

In a recent campaign for a national health and wellness brand, Butler/Till partnered with SWYM.ai and OpenX to rethink where programmatic optimization should happen — not after impressions are served, but upstream, before the bid.

The results were clear: materially stronger performance across 18 U.S. markets, driven not by more aggressive bidding, but by smarter supply selection. Instead of reacting to outcomes after the fact, the buy was shaped at the source, determining which inventory and supply paths were allowed into the workflow in the first place.

To unpack how that upstream approach works in practice, we spoke with Andrew Altersohn, Co-Founder and President of SWYM.ai. He explains how AI-native decisioning evaluates supply in real time, why shaping supply before the bid changes outcomes, and what this shift means for the future of programmatic buying.

1. How does SWYM’s AI-native platform differ from traditional optimization approaches in programmatic buying?

SWYM’s platform differs fundamentally from traditional DSP optimizations by making decisions upstream rather than post-bid. Instead of relying solely on historical conversion data after impressions occur, SWYM evaluates inventory and supply paths in real time to identify the most relevant, high-performing opportunities before the bid. 

This approach allows for smart supply, cleaner bidstreams, and improved efficiency, driving better KPIs like CPA, CPC, and ROAS, while also reducing the number of SSPs and domains involved.

2. Can you describe how your decisioning layer evaluates supply paths in real time, and what made OpenXSelect the right partner for activation?

SWYM’s decisioning layer continuously analyzes bid requests, local campaign performance, audience targeting signals, and brand safety data across all markets. It dynamically prioritizes supply paths that are most likely to deliver desired outcomes, effectively pre-selecting quality inventory for activation. 

OpenXSelect was an ideal partner because it allowed this upstream curation to seamlessly integrate into the buy-side workflow, giving access to premium, transparent inventory and leveraging OpenX’s identity graph to scale campaigns efficiently across 18 U.S. markets.

3. What role did feedback loops play in improving performance across the 18 markets?

Feedback loops were critical. SWYM continuously ingested local campaign performance and audience data to refine supply path selection in near real time. This learning mechanism allowed the platform to identify which bid requests were most likely to drive actions in each hyper-local market, resulting in significant efficiency gains — such as a 57% reduction in media CPA — and improved optimization over the campaign’s lifetime.

4. How does supply side curation (before the bid) change outcomes compared to standard DSP-side optimizations?

Upstream supply shaping gives SWYM control over what inventory even reaches the DSP, rather than reacting to outcomes after the bid. This increases the number of relevant bids to evaluate while eliminating duplication and waste, and reducing exposure to low-quality or non-performing domains.  Ultimately, dynamic supply shaping ensures that each bid has a higher likelihood of success. In this case, it translated to fewer SSPs and domains, lower CPMs, better CPCs, and a more predictable and efficient path to meeting KPIs.

5. How did the reduction in SSPs and domains impact efficiency and brand safety for the client?

By cutting 91% of SSPs and 37% of domains, SWYM reduced noise in the supply path, giving buyers fewer, higher-quality opportunities to bid on. This simplification improved brand safety by limiting duplicative bids and overlapping supply paths while reducing exposure to non-premium or risky inventory and increased efficiency by lowering CPMs and CPCs. The streamlined supply chain also enhanced SWYM’s ability to match audiences at the hyper-local level, directly contributing to the 57% lower CPA achieved across 18 markets.

6. From your perspective, what does this case study demonstrate about the future of AI-powered supply curation?

This case shows that AI-powered supply shaping is the future of programmatic buying. It demonstrates that precision, efficiency, and scale are not mutually exclusive: with upstream, intelligent decisioning and real-time learning, advertisers can execute hyper-local campaigns at national scale while reducing waste, improving performance, and maintaining brand safety. 

The ability to curate supply before the bid, combined with robust feedback loops, points to a new era where programmatic success is defined not just by how you bid, but by what you bid on.