Newton Research's Unlimited Analytics agents lift Horizon Media's ROAS 10-20%
Unlimited Analytics, Newton Research's new agentic layer, went live Tuesday with agents that handle campaign planning, activation and measurement inside one system. The causal modeling at its core lets a client type a question like what happens if we shift 20% of budget from social to programmatic and get a prediction built from millions of data points on creative exposure, timing and audience. Horizon Media has run the tool for about a year inside its Horizon Blu platform, pulling in bid, creative and geographic data it didn't previously have access to, and has seen its return on ad spend rise 10% to 20%. Newton CEO John Hoctor said data analysts typically spend about 80% of their time on data preparation, the work that comes before any real insight, before they can focus on measurement and client strategy.
Horizon Media used the causal models to find that a client's custom health-conscious audience segment converted at twice the average incremental rate on CTV, driven by creative featuring high-protein breakfasts and store-location calls to action, and shifted the client's budget toward that message. Agencies running agentic layers like Unlimited Analytics get ID-level visibility into which touchpoints drove a given conversion, freeing analysts from data assembly so they can focus on interpreting results and advising clients on where budget should move next. Newton connects to this data directly through Snowflake, keeping everything in one workflow, an approach Horizon says reduces the chance signals slip through the cracks as agencies refine their own marketing mix models.