Last-touch attribution is often criticized for misallocating budgets and stalling customer growth.
Vlad Chubakov of Delve Deeper advocates for a new ad measurement model using platform data, incrementality studies, and MMMs for clarity and control.
Google’s DV360 platform sees updates, enhancing YouTube and CTV strategies, but challenges with inventory quality remain.
Chubakov calls for AI integration in DSPs to provide immediate insights, noting current tools lack contextual understanding.
Ad measurement is broken. Clicks lie, conversions mislead, and last-touch attribution rewards the wrong channels. The fix is a measurement trifecta—platform data to optimize, incrementality to prove lift, and MMMs to see the full picture.
Vlad Chubakov, Associate Director of Programmatic at performance media agency Delve Deeper and the voice behind Programmatic 101 (also on X), pushes a new playbook. He advocates rebuilding measurement with clarity and control at the core.
The ultimate trifecta: “We have platform metrics, we have some incrementality studies, we have MMMs—and ideally those three things should be interconnected with each other,” says Chubakov. In this model, “platform data should be used for tactical optimization so you can allocate budget between different line items, and experiments should be used for understanding incrementality of your formats,” he explains. The approach reflects a broader shift toward more rigorous, privacy-resilient measurement.
Mind the gap: The need for a stronger measurement framework is clear. “Only 40% of web is addressable,” says Chubakov. “That means 60% of users don’t have cookies or other identifiers.” With so much signal loss, he warns, “platform conversions is not a very accurate way to measure your campaign performance. Especially for CTV, for audio, and for video. There’s a big challenge,” Chubakov says.
Last-touch lie: The problem runs deeper than signal loss. “Our industry is heavily reliant on last-touch attribution,” says Chubakov, calling it “a big problem.” The model, he explains, “creates a situation where bottom-of-the-funnel channels are oversaturated with budgets and unable to drive performance,” ultimately limiting new customer growth and stalling scale.
360 rotation: DV360 is finally moving. “For the last seven years almost, there were no major updates to the platform,” says Chubakov. But recently, Google has rolled out notable changes, especially for YouTube and CTV. DV360 is “not the best choice for open market campaigns,” he notes, but a strong contender for YouTube-first strategies. It remains the most widely-used DSP, with a clean interface and direct YouTube access marketers can’t ignore. Still, Chubakov warns, “there is a challenge with inventory quality and with open web capabilities,” so success depends on careful execution.
Intelligence > automation: Chubakov sees a clear role for AI inside DSPs. “I want LLMs to be built in the DSP. It would allow buyers to pull reports and receive immediate insights that can be implemented right away,” he explains. But today’s tools fall short. “Google has a chatbot for reporting, but it doesn’t understand the context of the campaign. So that’s a real hurdle,” says Chubakov. It’s intelligence, not just automation, that actually moves the needle.
© Copyright 2025 State of Streaming
