Data Infrastructure Review

ExchangeWire Data Infrastructure Review

The data-driven marketing data stack is a complex stew of TLAs and overlapping technology. From a publisher and marketer perspective it can be difficult to discern what technology matters most for day-to-day functions.

At a high level, there are four key layers to look at: onboarding; scaled data segmentation; activation; and a robust underlying data privacy strategy.

Onboarding and aggregating the signals

The industry has given rise to a range of identity solutions, designed to enable targeted marketing and advertising in a world where third-party cookies and device identifiers are no longer available. Context, for instance, is now an important go-to targeting signal for the industry, adding to a growing range of ‘signal ingredients’.

Onboarding and optimising these data sources is a critical function as fragmentation increases. Capabilities continue to overlap between traditional onboarders, CDPs and bespoke first-party data solutions.

Data intelligence and making sense of the signals

Getting a single view of your customer is the holy grail of modern-day marketing. There are multiple solutions vying to be the “data hub” for interpreting the mess of signals that both publishers and marketers aggregate. Marketing cloud businesses have invested heavily in this area, as the “intelligent data layer” evolves beyond the clunky CRM.

The CDPs (independent and marketing cloud) have been pushing the boundaries around the optimisation and segmentation of onboarded data sources. But things in this industry are never that simple: CDPs are increasingly being used as a privacy-first solution rather than an all-out replacement for traditional CRMs. There is still lots of innovation to come in this space.

Data segmentation

It is important to understand that segmentation will never be modular. But for the purposes of this overview, we will focus on two areas:

First-party data segmentation: Many publishers and marketers will activate their own first-party segments. Vendors are already working across various layers (onboarding, data intelligence, segmentation and activation). This is where the model bleeds into one, as technology performs multiple functions that adhere to privacy laws and platform data use.

Data segmentation at scale: In a data-privacy world, clean rooms will become the go-to technology to scale data segmentation. Publishers, marketers and agencies will need this tech to continue programmatic trading. There will be many clean room offerings: horizontal, vertical and regionalised. The key to success in this vertical will not necessarily be the tech - as it is not that difficult to spin up. The differentiation will be the integrations with data providers - marketers, publishers and agencies. All will need to be interoperable to maximise performance.

Data activation: how to get your data into the real world?

It’s all well and good bunkering data in a clean room or on an AWS instance. But it really has no value being siloed in an online database. The plumbing between a clean room and your chosen media channel is crucial in the future data stack. ID graphs will likely be a key function in this area, as marketers and publishers look to get “joined data” and first-party data into their chosen media channel. This activation will be dictated by privacy regulation and platform (Apple and Google) constraints.

A robust data privacy strategy must underpin the data stack

For the consumer, control and protection needs to be at its core. From onboarding to segmentation to activation, all will need to be underpinned by a strong data privacy strategy. Utility privacy tech, such as data governance, will need to be built into every step. There will also need to be proper opt-in and opt-out functionality throughout the process. If this is secondary, marketers and publishers will not be able to push-in, optimise or activate any type of data. You ALWAYS need to start and end with privacy.