The Trade Desk
AI copilot for optimizing programmatic campaigns (DSP)
In 2025, The Trade Desk's Kokai media-buying platform optimizes programmatic campaigns through its Koa AI: management cites -26% CPA, -58% cost per unique reach, and +94% CTR, with about 85% of clients using it by default in the third quarter of 2025, and a Specsavers campaign in the UK nearly halved its time to convert.
Key points
- Koa AI copilot optimizes bids, targeting, and cross-channel budget in the Kokai DSP.
- Real-time impression scoring across display, CTV, video, audio, retail media.
- Management figures Q3 2025: -26% CPA, -58% cost per reach, +94% CTR.
- Evidence A, confirmed status, about 85% of clients on Kokai at end of 2025.
Objective
Lower acquisition cost and extend the useful reach of programmatic campaigns by handing an AI the optimization of bids, targeting, and cross-channel budget allocation, rather than driving it by hand.
The deployment
Kokai is The Trade Desk's media-buying platform (DSP), released in 2023 and widely deployed through 2024. Its in-house AI, Koa, serves as a copilot to the media buyer: it scores ad impressions in real time, adjusts bids, allocates budget across display, CTV, video, audio, and retail media, and surfaces optimization levers. The buyer (brand or agency) keeps control of the strategy and guardrails; the AI handles the fine-grained steering across billions of impressions. The Trade Desk made the migration of its clients to Kokai the core of its 2024-2025 roadmap and reports on it at every quarterly result. In the UK, Specsavers and its agency Manning Gottlieb OMD ran four full-funnel omnichannel campaigns in Kokai combining video, CTV, and display. Two sets of aggregate figures coexist: the brand's official pages cite 43% lower cost per unique reach, 24% lower cost per click, and 27% lower CPA; at the third-quarter 2025 results, management put forward more recent gains of 26% on CPA, 58% on cost per unique reach, and 94% on CTR. Both series come from The Trade Desk and cover different scopes and measurement periods.
Results Proof A
The scaling and liveness rest on official financial results: Kokai adoption and revenue growth published in the 2025 quarterly results (primary source), aggregate performance gains put forward by management at the Q3 2025 earnings call, corroborated by an official quantified case study (Specsavers). Several consistent sources from The Trade Desk.
How it works
Inferred typical approachThe internal detail is not public. Here is a proven approach that leads to the same result, to adapt to your stack.
How it runs, concretely
For ops teams-
1Set objective and guardrails marketing / media trading (brand or agency)
The buyer sets the target KPI (CPA, reach, CTR), the budget, the channels, and the brand rules in Kokai.
-
2Connect data and measurement data team / agency
Connect first-party conversion signals, audiences, and where relevant retail media, so the AI optimizes toward the right event.
-
3Score and bid in real time AI (Koa)
Koa scores each impression, adjusts the bid, and allocates budget across display, CTV, video, and audio at each opportunity.
-
4Read the KPIs and arbitrate media trading / marketing
The team tracks CPA, reach, and CTR, adjusts the guardrails, and redirects media pressure toward what converts.
The first-party conversion and reach signals tied to the campaign (site events, bookings, purchases). Without a clean, reliable conversion signal, optimization toward CPA loses its target and falls back on less useful proxies.
How your customers perceive this type of use
Sourced studiesLe pricing algorithmique est le terrain le plus inflammable : 68% des consommateurs disent se sentir leses quand les marques utilisent le pricing dynamique et 80% jugent plus dignes de confiance les marques aux prix constants (Gartner, 2024). L'equite percue varie selon le secteur : le pricing dynamique n'est juge juste que par 33% a 40% des repondants selon qu'il s'agit de concerts ou de cinemas (YouGov, 17 marches). Le prix personnalise par les donnees individuelles est le plus rejete : 47% des Americains s'y opposent fermement (Consumer Reports, 2024).
Acceptance conditions
- La constance des prix comme signal de confiance : 80% jugent plus fiables les marques aux prix stables (Gartner 2024)
- Le secteur conditionne l'equite percue : le pricing dynamique est mieux tolere pour les cinemas (40% le jugent juste) que pour les concerts (33%) (YouGov 2024)
Red lines
- Le pricing dynamique percu comme abus : 68% se sentent leses (Gartner 2024)
- Le prix individualise a partir des donnees personnelles : 47% d'opposition ferme (Consumer Reports 2024)
- Les frais caches et hausses imprevues, vecus par 79% des consommateurs sur un an et associes a la perte de confiance (Gartner 2024)
Sources: Gartner 2024 · YouGov 2024 · Consumer Reports 2024
How to replicate
Inference, not sourcedData prerequisites
- A clean, reliable first-party conversion signal (site event, lead, purchase) to optimize on
- Activatable audiences or segments and a coherent reach measurement across channels
- A legal basis and GDPR consent for data collection and activation, a third-party cost key or an identifier like UID2 if the third-party cookie is missing
Org prerequisites
- A programmatic budget large enough for AI optimization to have material to learn from
- An in-house media trading team or an agency able to set objectives and guardrails and then read the KPIs
- Marketing / data agreement on the definition of the target conversion and the brand rules
Possible stack
- The Trade Desk Kokai with Koa AI
- Another DSP with an AI optimization layer (equivalent approach)
- A cross-channel conversion and reach measurement component
The plan, step by step
- Step 1Choose the conversion event to optimize and verify its tracking is cleanDeliverable: A reliable conversion signal connected to the platform
- Step 2Connect audiences, measurement, and cross-channel inventory in the DSPDeliverable: An omnichannel campaign set up with objectives and guardrails
- Step 3Let the AI score and bid, framing budget and brand rulesDeliverable: Real-time optimization toward the target KPI
- Step 4Read CPA, reach, and CTR by channel and reallocate media pressureDeliverable: A reading loop linking spend, conversions, and reach
First step: Verify that the target conversion event is correctly tracked and usable: that signal drives the whole AI optimization, and without it the CPA gains do not materialize.
The same pattern, proven in other industries
adidas
AI-driven audience prospecting and bidding (ARO) across the full funnel
Allwyn
AI-driven omnichannel programmatic buying with retail data activation
Nivea
genAI testing and optimization of creative elements + automated localization at scale
Sources
- S1 The Trade Desk Reports Second Quarter 2025 Financial Results Primary archive pending
- S2 Drive better outcomes with The Trade Desk's Kokai Interested party archive pending
- S3 Specsavers boosts bookings with AI-powered, full-funnel omnichannel strategy Interested party archive pending
- S4 Strong Kokai Adoption Fuels TTD's Revenue Pipeline & Growth Prospects (recap resultats T3 2025) Secondary archive pending
An error, newer info, a source?
This page lives on its accuracy. If a figure has moved, if the deployment has changed, or if you have a higher-quality source, tell us. Every sourced correction is verified before publication.