Databricks’ Jake LaDuke: Marketing is Shifting From an Attention Economy to a Prediction Economy
The way marketers prove value is changing. Attention metrics that have anchored advertising since its inception are giving way to something more demanding. It starts with the ability to predict business outcomes before a campaign even launches.
“CMOs need to be able to have a predictive layer of performance before they even green light something. This is shifting from a lagging indicator of performance into a leading indicator of performance,” Jake LaDuke, Global GTM Lead for Media, Entertainment and Advertising at Databricks, told Beet.TV contributor David Kaplan at the Beet.TV Leadership Summit with Omnicom Media during Advertising Week New York.
That shift carries direct implications for how CMOs talk to CFOs. That evolution also speaks to how AI investment gets justified inside organizations still learning to speak the language of prediction.
Agentic marketing in practice
A major pet retail chain illustrates what agentic transformation looks like when it actually works. By consolidating loyalty data on Databricks and deploying agents to orchestrate personalized messaging, the retailer moved away from rigid product and services calendars toward dynamic, data-driven outreach.
“They shifted away from a product calendar, a services calendar, into being able to harness all of their loyalty data and unleash agents to send out personalized messaging,” LaDuke said. “They’re seeing 22% incremental performance on some of their components and a 12% lift on transactions.”
CMOs need to look good to their CFOs
LaDuke frames the new accountability challenge in terms of the relationship between the CMO and CFO. The question a CFO increasingly asks is simple: given this budget, what business results should I expect?
“Part of what a CFO wants to understand is to be able to look across to their CMO and say, ‘If this is the budget, what can I expect as far as business results?’” LaDuke said. “You have to harness all the various sources of truth of what performance looks like across your media and advertising and orchestrate those into something that is a singular metric on what it did to the business.”
Four Cs underpin successful AI transformation
LaDuke distills the first principles behind any successful AI transformation into four areas. Context refers to how well agents can access and use the broader data estate. Control covers who (and what) has access to that data, including the agents themselves. Cost addresses the shift from token maximization to value maximization. Choice means having the right model for the right job.
“Between CMOs and CFOs, we’re seeing this shift from a world of token maxing into a world of value maxing,” LaDuke said.
Think big, aim small
The tension LaDuke leaves marketers with is also a practical framework. AI should be envisioned as a growth lever for the whole business. The use cases that unlock it, such as audience planning, activation, measurement orchestration, should be tackled incrementally.
“It’s thinking big about how AI can be that growth lever for your business and then aiming small when it comes to the use cases you need to unlock in order for that to be successful,” LaDuke said.
The way marketers prove value is changing. Attention metrics that have anchored advertising since its inception are giving way to something more demanding. It starts with the ability to predict business outcomes before a campaign even launches.
“CMOs need to be able to have a predictive layer of performance before they even green light something. This is shifting from a lagging indicator of performance into a leading indicator of performance,” Jake LaDuke, Global GTM Lead for Media, Entertainment and Advertising at Databricks, told Beet.TV contributor David Kaplan at the Beet.TV Leadership Summit with Omnicom Media during Advertising Week New York.
That shift carries direct implications for how CMOs talk to CFOs. That evolution also speaks to how AI investment gets justified inside organizations still learning to speak the language of prediction.
Agentic marketing in practice
A major pet retail chain illustrates what agentic transformation looks like when it actually works. By consolidating loyalty data on Databricks and deploying agents to orchestrate personalized messaging, the retailer moved away from rigid product and services calendars toward dynamic, data-driven outreach.
“They shifted away from a product calendar, a services calendar, into being able to harness all of their loyalty data and unleash agents to send out personalized messaging,” LaDuke said. “They’re seeing 22% incremental performance on some of their components and a 12% lift on transactions.”
CMOs need to look good to their CFOs
LaDuke frames the new accountability challenge in terms of the relationship between the CMO and CFO. The question a CFO increasingly asks is simple: given this budget, what business results should I expect?
“Part of what a CFO wants to understand is to be able to look across to their CMO and say, ‘If this is the budget, what can I expect as far as business results?’” LaDuke said. “You have to harness all the various sources of truth of what performance looks like across your media and advertising and orchestrate those into something that is a singular metric on what it did to the business.”
Four Cs underpin successful AI transformation
LaDuke distills the first principles behind any successful AI transformation into four areas. Context refers to how well agents can access and use the broader data estate. Control covers who (and what) has access to that data, including the agents themselves. Cost addresses the shift from token maximization to value maximization. Choice means having the right model for the right job.
“Between CMOs and CFOs, we’re seeing this shift from a world of token maxing into a world of value maxing,” LaDuke said.
Think big, aim small
The tension LaDuke leaves marketers with is also a practical framework. AI should be envisioned as a growth lever for the whole business. The use cases that unlock it, such as audience planning, activation, measurement orchestration, should be tackled incrementally.
“It’s thinking big about how AI can be that growth lever for your business and then aiming small when it comes to the use cases you need to unlock in order for that to be successful,” LaDuke said.
