When we think of privacy, the first thing that comes to mind for many of us is regulations such as the GDPR. But the law is not the only thing that has changed the rules of the game; technology such as ITP has forced new standards through changes such as full third party cookie blocking that have fundamentally altered the foundations that ad-tech was built on. As a direct result, internet users today have greater expectations than ever for companies to act responsibly and transparently when it comes to data. And as user expectations change, so must the companies who seek to interact with them online.
In parallel, there has been a sea change in the sophistication of machine learning in digital marketing platforms. Google and Facebook are now better than ever at automating optimisation through their algorithms.
How do digital marketers navigate this new world and what do the combined forces of privacy and smarter machine learning mean for the role of data in data-driven digital marketing? Here we set out three points to consider when setting out a strategy in the privacy-first era.
Privacy requires legal expertise, but it also has profound effects on user experience (there’s nothing worse than a bad cookie banner) and the way that we communicate with customers. At the same time, it has a significant impact on the ways that data can be used, setting a much higher bar for collection, analysis, and actionability.
A cross-functional privacy working group should be established with the goal of understanding the value of data to the organisation and championing customer-centric privacy. Unnecessary data need no longer be collected, and the value of useful data to the company should be clearly documented. The group also needs to understand how to comply with regulations, but ideally should set their ambitions higher and understand how to communicate clearly with users, protecting the voice of the brand and earning their trust.
In 2016, the dream for attribution was to build the perfect multi-touch model. Anybody who worked on such a project might remember that the dream was rarely realised. Now, in 2020, it is impossible. Technology developments such as ITP, an increasing growth in cross-device journeys, and the fact that consent is required before user interactions can be measured means that paths are more fragmented than ever.
Luckily, the attribution toolbox has also developed since 2016. Tools such as MMM (Media Mix Modelling) and Lift and Incrementality studies have come a long way and are much more readily available in the platforms we work with.
This topic will be explored in greater depth in part five of our series, but the headline is that marketers must embrace a multi-faceted approach to attribution. Path based models, MMM, and incrementality studies each have strengths and weaknesses, and our ability to combine the data from each method in the smartest way will give us the best chance of success when it comes to the correct setting of bids, budgets, and targets.
Due to privacy considerations, there is a higher bar for the data we collect and use, so it’s important to choose wisely. When it comes to the question of what data to capture to add value within the organisation, the machine learning revolution in the platforms is a core consideration. As Google and Facebook become smarter and smarter at using algorithms to optimise for outcomes, our role as digital marketers is to train those algorithms to perform to our advantage.
Take an e-commerce business for example. Few e-commerce websites operating successful paid search or social campaigns are using Total Sales as a conversion signal in 2020. Many, on the other hand, are still using Revenue. Is this the datapoint we want to train the platforms to use to maximise the impact of digital marketing on the business? Part of a successful data strategy in 2020 will be to find better signals. For example, optimising towards outcomes such as profit, new customer acquisition, or lifetime value, and doing so in a privacy-first and automated way. Explore more on this topic in part four of our series.
Leading the way
Whether your company has already made a lot of progress, or whether you are still at the beginning of the road, a cross-functional working group is essential to developing a strategy that balances the need to create value from data within the organisation, the need for customer-centric communication, and exceeding expectations when it comes to privacy. Those who move first stand to gain the most, both in terms of the trust of their users as well as the ability to earn a competitive advantage by creating processes and structures that drive excellent results in a privacy-first way.