The industry agrees that relevance is the jewel in the crown of effective advertising, with first-party data an important alloy to hold it all together. In this piece, Fritz Richter, CTO and co-founder, adsquare, (pictured below) looks at the requirements for a secure data marketplace and suggests five key elements.
Advertising has become a battlefield for relevance. The rise of the empowered consumer, particularly digital natives who grew up 'mobile-first', has turned up the pressure on brands and marketers to deliver personal communications aligned with context to ensure they approach the right audience, in the right moment, with the right message.
To do this effectively requires advertisers, and their agencies, to master two capabilities. First, they must become 'programmatic pros'. This assumes advertisers have sophisticated tools and a clear understanding of how they must use these complex real-time platforms in order to address and engage their target audience. Second, advertisers and agencies must grasp the core importance of data and dedicate efforts to fuel their platform stacks with data (static and dynamic) around audiences, locations, and context.
But no one goes it alone. Delivering highly-valuable, highly-relevant, mobile marketing and communications to today’s mobile-first audience, in the appropriate context, also requires access to data from a variety of companies – including app developers, publishers, telcos, and retailers. Their data, also known as 'first-party data', is highly valuable because these data owners have a direct relationship with their customers. The information and insights they have access to, as a result, add tremendous depth and meaning to audience targeting, depending on where the data owner sits in the consumer journey.
While there are clear benefits to first-party data, concerns about data security, fuelled by a rise in reports about data leakage and the misuse of data ownership, are the bottleneck. To overcome this hurdle, the industry must address privacy and security concerns head-on, and take concrete steps to give data owners what they need in order to activate (and monetise) their data assets for programmatic advertising.
These challenges drive industry demand for a new kind of marketplace for data, one that ensures a secure environment for first-party data. Attention to the following five key features will guarantee this platform offers all stakeholders a secure environment.
1. Work with a secure and state-of-the-art server environment
A server environment for first-party data should be built security-first, which means that it would need to be made up of servers that are protected against internal and external attacks. These servers should also only process encrypted data and restrict access to raw data. To work with data from multiple providers, these services would also need to comprise multiple dedicated 'data silos', which would correspond to each individual data provider.
2. Give data owners full control with private deals
In the media world, it took a long time for publishers to make their premium inventory accessible for programmatic buyers. The catalyst in this case was the establishment of so-called 'private marketplaces' (PMPs). These PMPs gave publishers supply-side instruments (and 100% control) regarding who had access to their inventory and under which conditions. Adopting the same approach for data would result in a controlled and transparent environment for first-party data. Data owners could then set rules around access (i.e. which segments are visible to which advertisers) and define the pricing based on the campaign and buyer.
3. Enable pre-bid enrichment so that data never leaves the platform in bulk
Traditional data platforms, as we know them from the cookie space, activate data for buying platforms via bulk uploads, including raw IDs. This has two downsides: First, dynamic data is not accessible in real-time; and second, the data platform makes all data available in one transfer. (It’s a bit like leaving your flat door open and posting a banner that says “Please do not rob me” in order to deter burglars. Once the door is open, what real good can this do?) This is why it is important to work with a data marketplace that has deep pre-bid integrations into DSPs, thus allowing data to be used only for enrichment.
In the pre-bid integration, the platform allows users to see only the meta information during the modelling step, thus barring access that would allow users to extract or download capabilities of the actual content. This, in turn, ensures that data is not abused. Instead, data is used only to enrich a specific bid request once, and then discarded. Moreover, the pre-bid integration allows the DSP to only see encrypted audience IDs, not the actual meta-information, such as the used attributes, or the audience. The final benefit of this model is efficiency, as it enables buying platforms to perform real-time decision-making and bidding adjustments based on dynamic data.
4. Make combining data points a requirement
Today, holistic audiences are built by combining a vast array of multidimensional data. In this process, more data is better; and it’s not sufficient to use just one single attribute as an audience. It’s why it makes sense for data providers to make their raw data available – and why using this data, in combination with other data, enables significantly better audience targeting. However, it’s also important to ensure that the raw data that contributed to the improved audience targeting can, under no condition, be 'reverse engineered' to determine which data provider (and which of their data assets) made this possible. The ability to draw a direct connection between the data owner and the data they contribute must be avoided. This can be achieved by a model that effectively blurs and crunches the data, combining data points for holistic audience targeting, but not exposing details that could trace the data to the data owner.
5. Put binding data usage agreements in place
Ensuring that data is processed in a controlled and transparent way is essential to establishing a secure environment built on trust. But it doesn’t stop there. The environment must also be supported by commitment, and clear legal contracts between data buyers and data owners must be in place, laying down exactly how, and for which purpose, data can be used. A secure data marketplace model must also ensure, through clear communication, that breaking these legally binding contracts is no trivial offense.
Follow these five steps and you will come to the conclusion that a data-driven and real-time ecosystem demands a new breed of data exchange. One that has been built mobile-first, works in real-time, and protects data ownership. Done right, first-party data, which is very valuable for advertising, can be properly and effectively harnessed to power holistic audience targeting and moment marketing. The result is more relevant advertising and a better outcome for all.