To learn more about The Drum data, in The New Data and Privacy Playbook, OMD’s Miles Pritchard expresses his optimism that machine learning-enhanced media personalization will reinvent a cookie-free future. sharing.
Google’s plans to remove third-party cookies have set the marketing world on fire for the last few years. Advertisers, publishers, and ad tech companies are all phasing out cookies in favor of alternative solutions that deliver the personalization that modern marketers love, at the scale they expect.
As alternative solutions are explored, machine learning is gaining traction and promises to enable the next generation of media personalization through a forward-looking approach to audience development, translation and activation.
Simplify machine learning
Machine learning is a computer system that uses statistical models to adapt, analyze patterns in data, and draw inferences from it.
This means using cloud computing to process massive amounts of historical data about audience demographics, media performance, conversion paths, customer lifecycles, etc., to make predictions about who to target and what media to place where. means that you can generate These forecasts are subject to change automatically to reflect current conditions.
Importantly, many data inputs and outputs do not require cookies and can be activated without an ID. Necessity is the mother of invention and the future of cookie-based audience segmentation and activation looks bleak, so the industry has been looking for solutions.
Our mission is to find ways to analyze and model data (cookies or otherwise) to extract meaningful insights, then translate and apply those insights into the context of media activation for personalization, relevance, and finding a way to provide information. performance.
Drive decision-making with simulation
One of the key drivers of machine learning-driven decision-making is the role simulation plays.
With simulation, you can analyze millions of potential future scenarios in minutes, including budget allocation, creative evolution, content personalization, bid evaluation, and more, to evaluate optimal outcomes. Together with other use cases, it has revolutionized many processes in media planning and buying.
Over the past few months, OMD has used machine learning in some interesting ways to create cookie-free marketing outcomes.
1. Trend modeling solution
These are solutions that leverage machine learning along with aggregate media.Leverage party web analytics data and customer records to generate propensity models that predict conversion/sales likelihood for attributes such as time, date, location, device, browser and content.
2. Natural Language Processing (NLP) Tools
This includes using NLP tools to translate audience profiles and map them to the most relevant audience taxonomies within the walled garden solution and contextual audience partners and publishers across open exchanges. will be
3. Custom Bid Optimization (CBO)
The CBO algorithm builds on propensity modeling and other inputs to generate automated bid metrics. This allows you to read, interpret, and act on impression-level signals in real-time to determine appropriate bid evaluations that reflect the latest user behavior and conversion paths.
new normal
Every system has a limiting factor. The main limiting factors in machine learning are the quality of the data that enters the system, the assumptions people make when building models, and unknown entities that are not captured within the modeling process.
We see all three of these challenges combined during the COVID-19 pandemic, with devastating impacts such as atypical trading patterns, supply disruptions and changes in consumer behavior. I saw it.
Today, with the digital footprint generated by our consumers, we have an abundance of raw data at our disposal. Machine learning has made it possible to clean, reconcile, and use this vast data resource to diagnose trends, derive insights, and generate predictive analytics.
Somewhere along the line, we stopped thinking of machine learning as a tool to build for the future. We started using it to build today’s solutions, ushering in a new generation of data-informed personalization solutions that have become today’s mainstream marketing tools.
If you want to read more about The Drum’s latest Deep Dive demystifying data and privacy for marketers in 2023, visit our special hub.
