Taking first-party data to the next level with machine learning

Over the past few years, data privacy has emerged as the major recurrent theme in ad tech. Data privacy regulations have both grown in number and strengthened in rigor, as regulators seek to put consumers back in control of their personal data.

While data protection laws have been around since the mid-20^th century, Europe started the ball rolling on the current era of more digitally-focused regulations with the introduction of the General Data Protection Regulation (GDPR) in 2018, followed by the California Consumer Privacy Act in 2020, and numerous other laws in different parts of the world. The European Digital Markets Act, which came into force in 2022, aims to further regulate the data gathering and sharing activities of big tech players such as Apple and Google, while these companies have also pitched in with their own data privacy initiatives. In 2021, Apple launched its App Tracking Transparency framework, which gives Apple users the right to opt out of being tracked across apps and websites they use, while Google is planning something similar as part of its Privacy Sandbox initiative.

At the same time, regulators have sought to tighten existing laws. Late last year, for example, the Federal Communications Commission sought to amend the Children’s Online Privacy Protection Act (COPPA) to make it harder for tech companies to collect and monetize children’s data.

These moves have impacted ad tech players in two ways. Firstly, it is now incumbent on any company involved in digital advertising to ensure that they meet all relevant standards in terms of how they collect, secure, manage, and process consumer data, and ensure that it is compliant with all relevant legislation. Key among these standards are SOC 2 Type 2 Certification; IAB TCF Membership; IAB OM SDK Recertification; IAB Tech Lab Standards adoption; CCPA Validation; GDPR compliance; and ISO/IEC 27001:2013 Certification.
Secondly, more and more companies have come to appreciate the power of first-party data and started to leverage this to run better-targeted, more personalized campaigns. First-party data is indeed a powerful asset, offering companies real insight into who their users are, what they are interested in, what they do, and what they buy.

But first-party data alone is only half the story. It is at its most potent when it is combined with other data points, but the problem is that these are often fragmented, residing on different platforms, and not easily able to be gathered into one dashboard, where they can be viewed holistically and acted on.

This is where machine learning comes in. When first-party data is combined with machine learning and integrated with other signals such as those coming from Mobile Measurement Partners, you get a really detailed picture of who your users are and what makes them tick.

Machine learning models can map user behavior across time, and across multiple channels and touchpoints, identifying emerging trends and shifts in user preferences. The combination of first-party data with real-time engagement data such as user app requests, ad interactions, anonymized in-app behavior data and attribution information allows for the granular segmentation of audiences, enabling the creation of multiple segments of users who share similar characteristics. Advanced ML technology such as deep learning can also optimize user profiles by understanding patterns from datasets including images and videos.

The data can also be used to create a model that analyzes the relevance of the data, allocates appropriate weight to its importance, and offers a predictive view of which ads will resonate best with the user, enabling you to spend your marketing dollars more effectively. These models can also help with remarketing, which offers a cost-effective way to reduce churn, boost retention, and maximize lifetime value by targeting a subset of “warm” users who have previously shown an interest in your product.

At the same time as creating multiple segments of users, machine learning also enables the rapid and easy creation of hundreds, or even thousands, of highly personalized and contextual creatives that are optimized for each segment and for different regions, using technology such as Dynamic Creative Optimization, which can quickly iterate and optimize elements of the ad creative for better engagement and conversion. All while adhering to the most stringent privacy regulations.

Machine learning also offers the capability to deliver analytical insights not only post-campaign but also in real-time, meaning optimization will be more dynamic. Marketers should be considering predictive analytics that incorporate both historical and real-time data in order to allocate ad budgets more strategically and to the channels with the highest potential for ROI.

I believe that this combination of first-party data and machine learning will help to redress the balance between the open internet and the walled gardens with respect to ad budgets. According to Trade Desk Intelligence and Canvas8, as of April 2023, US consumers spent 59% of their online time on the open internet, and just 41% in the walled gardens of Google, Meta and Amazon. Despite this, the walled gardens attract 52% of ad spend, compared to just 48% for the open internet, due to the walled gardens’ targeting, optimization and reporting tools. If the wider use of first-party data and machine learning enables the open internet to offer the same level of targeting, personalization, accountability and measurability as the walled gardens, then more ad spend will head that way, which can only be a good thing for those who believe in the power, and the value, of the open, ad-funded internet.

Based in Silicon Valley, Greg has contributed over two decades of invaluable digital advertising expertise to the company. Throughout his distinguished career, Greg has occupied significant roles top organizations such as Yahoo and AOL, sharpening his acumen in the digital advertising sector. Before joining Mobvista, he oversaw the programmatic exchange at Amobee and spearheaded programmatic partnerships at Celtra.