Overview
For years, programmatic advertising has helped media buyers optimize audiences, bids, and creative while treating media quality as something to measure after the fact rather than a variable in bidding decisions. A leading global hotel brand and its agency partner dentsu decided to challenge that assumption.
Operating thousands of hotels across more than 100 countries, the brand invests heavily in digital media to drive qualified traffic and bookings across competitive travel markets in Europe. After building a sophisticated programmatic operation with dentsu, incremental gains became harder to find.
Traditional optimization tools offered limited visibility into page-level quality before bidding, leaving campaigns exposed to low-attention environments, excessive ad clutter, and poor consumer experiences. The team needed a way to distinguish inventory that was technically viewable from inventory that could genuinely earn a traveler’s attention, all before placing a bid.
With that challenge in mind, the brand and dentsu set out to:
- Reduce cost per booking
- Increase hotel room bookings
- Eliminate waste from low-quality inventory
- Validate that AI-powered media quality scoring could improve business outcomes without increasing media costs
Ultimately, the team wanted to prove that selecting higher-quality impression opportunities on the sell side, before bidding, could improve programmatic efficiency at scale. Doing so would require that media quality become a real-time, pre-bid decisioning signal rather than a reporting metric.
Solution
To turn media quality into a buying signal, dentsu activated Chalice’s custom decisioning through Index Marketplaces. Within the Marketplaces platform, deals are curated and built against targeting keys, and when an impression matches, Marketplaces creates the Deal ID that makes it buyable. Chalice supplied one of those keys: a real-time, campaign-specific decision on whether an individual impression cleared the quality bar dentsu had set.
Chalice dynamically adapted bidding decisions based on the brand’s campaign objectives and the quality of each individual opportunity. The model scored every ad opportunity across more than 35 signals, including cluttered layouts, high refresh rates, and poor content ratios, continuously reprioritized opportunities and adjusted bid values in real time as the campaign ran.
Chalice ran that model inside Index Cloud, which lets partners run their own models, data, and applications inside Index infrastructure at the point where impressions originate, rather than hosting them in outside infrastructure and waiting for bid requests to arrive.
That distinction shapes what a model can do. A model running outside the exchange only ever sees a fraction of what is available: fewer impressions, and less detail about each one. The signals that matter most to a quality decision are often the first ones lost.
Running inside Index Cloud removed those constraints. Chalice scored roughly half a billion webpages rather than a sample, and ran hundreds of millions of bidding instructions, enough to express the full quality model instead of a compressed version. For a campaign that came down to separating genuinely high-quality opportunities from marginally viewable ones at scale, that expanded coverage is what made the difference.
By evaluating opportunities before the bid rather than measuring quality afterward, dentsu, Chalice, and Index transformed media quality into a real-time optimization signal instead of a retrospective reporting metric.
“What Chalice did here is what we built Index Cloud for. A partner brings their own model and data, runs closer to where impressions originate, and the buyer sees it in the results. The same approach can unlock new possibilities across a range of use cases.”
Michael Richardson, VP of Product
Index Exchange
Results
The campaign proved that media quality is a measurable performance lever, not just a brand safety consideration.
Over the two-month campaign in Europe, the brand achieved:
- 41% Lower cost per booking
- 14% More hotel room bookings
Eliminating low-value impressions before bidding allowed the brand to direct more spend toward high-quality environments built for engagement and conversion. Chalice continuously evaluated and reprioritized every ad opportunity in real time, adapting as inventory and conditions changed rather than relying on static rules.
“The industry has treated media quality as a reporting exercise for too long. What made this campaign different was running Chalice’s custom decisioning directly inside Index Cloud from day one, scoring close to half a billion pages before the bid rather than measuring quality after. That’s the architecture that made the 41% reduction in cost per booking possible. It’s what happens when AI has the right signals at the right moment in the transaction.”
Freddie Turner, Managing Director EMEA
Chalice AI
The results proved that when media quality informs sell-side decisioning, it can directly improve performance and deliver business outcomes. Together, dentsu’s media strategy, Chalice’s AI-powered curation, and Index’s sell-side infrastructure showed how applying intelligence earlier in the transaction helps buyers make better decisions before every bid.
“Our goal is always to find smarter ways to turn media investment into meaningful business results for our clients. This campaign showed that evaluating media quality before the bid can materially improve performance, helping us reduce wasted spend while driving more bookings. It’s a strong example of how sell-side decisioning and more intelligent programmatic infrastructure can create measurable value for marketers.”
John Thankamony, Managing Director, Total Addressable
dentsu
See how you can bring your intelligence directly into the exchange through Index Cloud, unlocking richer signals, greater scale, and faster decisioning.
