This post was last updated on October 8, 2026.
Overview
Retailers are under increasing pressure to make every media dollar work harder, especially as e-commerce traffic becomes more competitive and less predictable. Many brands still rely solely on demand-side optimization, which means they aren’t taking advantage of valuable upstream sell-side signals—the signals that reveal which impressions are most likely to generate site visits.
This was the case for a major apparel retailer aiming to drive more high-intent shoppers to its website through always-on programmatic online video (OLV) campaigns.
To improve both scale and cost per site visit (CPSV), the retailer partnered with a major agency holding company and inPowered AI to explore how sell-side decisioning through Index Marketplaces could drive more efficient site visits.
By applying predictive intelligence earlier in the supply path, the team aimed to prioritize high-value impression opportunities and route spend toward those supply moments most likely to drive action.
Solution
inPowered AI built a custom predictive model using the retailer’s historical DSP performance data. Purpose-built for this always-on campaign, the model scored each impression based on its likelihood to generate a site visit, using patterns observed across the retailer’s past campaigns.
inPowered AI deployed that model as a container inside Index Cloud, a neutral compute environment inside the exchange. Partners run their own models and data there, at the point where impressions originate, rather than hosting them elsewhere and waiting for bid requests to arrive. Running inside Index Cloud, the model evaluated all relevant sell-side signals in real time, including contextual and geographic factors such as time of day, page URL, browser version, operating system, carrier, and zip code.
When the model identified high-potential impressions, it applied a unique deal ID so spend could automatically route toward those opportunities.
Hosting the model directly within Index’s infrastructure allowed inPowered AI to apply predictive sell-side intelligence before bids were ever sent to the DSP, unlocking performance gains that aren’t possible when optimization occurs only downstream.
“Deploying our models inside Index’s infrastructure is what makes this approach possible. Index Cloud lets us decision across the full supply landscape, which leads to more consistent and meaningful performance gains.”
Peyman Nilforoush, CEO and Co-Founder
inPowered AI
Results
By applying predictive intelligence at the supply layer, the retailer achieved a major improvement in site-visit efficiency: 75% reduction in cost per site visit (CPSV).
Instead of relying exclusively on downstream signals, the team used real-time supply signals to prioritize impressions with a higher likelihood of delivering outcomes—cutting waste, increasing performance, and making every media dollar work harder.
“This is a clear example of what sell-side decisioning can unlock. Brands that only optimize on the buy side are missing the broader opportunity. By evaluating the full open internet through Index Cloud, we’re seeing stronger results across nearly every KPI we optimize toward.”
Pirouz Nilforoush, President and Co-Founder
inPowered AI
See how you can bring your intelligence directly into the exchange through Index Cloud, unlocking richer signals, greater scale, and faster decisioning.
