营销数据智能突破：用于贝叶斯M​​MM和消费者测试的Agentic系统.pdf

1、Powering Marketing Science on DatabricksAgents for Bayesian MMM and Consumer TestingDr.Luca FiaschiDatabricks Data+AI Summit 2025Dr.Luca FiaschiPartner PyMC LabsFormer:CDAO Mistplay VP ML Stitch Fix VP Data&ML HelloFresh Director AdTech Rocket Internet SEDoingAI Masters Marketing Execution,Strategy

2、Lags3AI is revolutionizing doing things.Now,it must evolve to revolutionize thinking.71%of marketeers use AI tools weekly to generate new contentAI personalization boost AOV 40%98%of AI-using sales team report strong efficiency improvements Thinking61%of CMOs report that their plans lack direction f

3、rom customer insights70%of CMOs face significant challenges proving campaign ROIMarketing Decisions:Bottlenecked By DesignThe Expertise GapThe Iteration LagThe Translation Barrier4Why does translating data into actions remains a major hurdle?Nearly 76%of data/AI professionals reports a talent shorta

4、geSophisticated Models(MMM,LTV)are slow to build and iterate;often only run 1-2 x yearFewer than half of data leaders feel their team effectively add valueThe“Gut Feel”ApproachKey Problems Why AI Agents Solve These Exact BottlenecksUnlock advanced modeling for the whole team its like a data science

5、lead on every project.Shift from monthly reports to real-time“what-ifs”for faster,smarter decisions.Democratizing ExpertiseAccelerating IterationTransform complex data into clear business language that drives action.Simplifying Communication56 Rapidly validates concepts and refines brand strategy wi

6、th synthetic consumer panels.Simulates consumer behavior for faster,data-driven product decisions.Transforms complex Media Mix Modeling into an agile,conversational process.From data prep to strategic insight delivery,end-to-end.Two Key Open-Source Examples:These Agents act as AI-partners to augment