Alexandra Theriault, Chief Growth Officer, Lotame, explains how brands that can activate their customer data across the channels where their audiences live will command a massive mindshare advantage.
Despite brands being years into the digitisation push, onboarding customer data remains a hurdle for many, compounded by poor match rates and the questionable methodologies connecting dots in a privacy-first world. Those with deep historical records have particular difficulty, and must design onboarding strategies around building digital touchpoints with customers new and old to avoid being left behind.
What is data onboarding and how does it work?
Data onboarding is the process of linking offline data – most commonly data sets in customer relationship management (CRM) platforms – with online identifiers to create a unified customer profile. CRM data might include email addresses, phone numbers and postal addresses, while online identifiers could be anything from an IP address or a first-party cookie to a mobile ad ID (MAID) or a universal ID.
Through this process, marketers can gain a “joined-up” view of their customer journeys as well as the behaviour of their target audiences across various offline and online channels – vital in a world where customers shop both in store and online and split their attention across multiple devices and media platforms. The ballooning complexity of omnichannel campaigns and anything but linear consumer journeys have made a holistic view more important than ever.
Offline data can be acquired from data marketplaces in the form of zero-party data or second-party data, but typically onboarding involves matching the first-party data a brand holds on its customers. Having a robust onboarding strategy has become particularly important as the gradual – and soon to be complete – deprecation of third-party cookies has limited the ease with which a brand can track its customers across the internet.
Data matching can be achieved through two different approaches: probabilistic or deterministic. Which is more appropriate depends on the quality and quantity of available data, the sites or devices that are being linked to, and the objective of the onboarding. Data onboarding with probabilistic matching uses AI/machine learning algorithmic models to weigh the likelihood that offline consumer data is representative of its online counterpart, while onboarding deterministically involves identifying an exact one-to-one match between offline and online profiles.
Both approaches are often used in tandem, as many CRM data sets are not large enough to be matched deterministically to online audiences at scale and even in the best-case scenario match rates for email addresses (which can become somewhat stale and unreliable as an identifier) hover around 50 per cent; a shortcoming that can be compensated for by layering in probabilistic matching.
What are the challenges of data onboarding?
For a brand to be able to pursue an onboarding strategy, first of all it needs data to onboard. Those with direct relationships with customers can incentivise the consented sharing of personally identifiable information (PII) through various touchpoints such as account registration, loyalty schemes or subscriptions. Marketers must consider how valuable and actionable this data will be to the brand. For example, running a competition might gather plenty of email addresses, but the records won’t necessarily be representative of the brand’s target customers.
Even brands that have customer data often struggle to activate it because it’s scattered across various departments, from customer services and ecommerce to marketing and media buying. Breaking down data silos and working from a unified customer data platform (CDP) ensures that data is accessible to everyone within an organisation. This process comes with challenges of its own as data sets may require transformation so that they are consistently formatted and duplications can be filtered out — a process that has been made simpler in recent years thanks to machine learning.
Then there is privacy to consider. Regulations such as GDPR and CPRA require that customers must provide informed consent for their data to be used. When PII is matched to a third party’s online identifiers it must be anonymised in the process and any behavioural data generated cannot then be de-anonymised and linked back to the customer. This is where broader analysis of audiences comes in, allowing behaviours to be tracked at the segment or cohort level without directly identifying the individual.
How does data onboarding provide an advantage in omnichannel marketing?
Successful data onboarding gives brands and the agencies supporting them significant advantages across a range of channels.
Within a brand’s owned properties and channels, data onboarding empowers the delivery of highly tailored and engaging digital experiences. By utilising the data onboarded into personalisation platforms, brands can customise website content, offers and messaging in real-time, and monitor its effects on customer relationships, engagement and conversion rates.
In digital display advertising, data onboarding streamlines the transmission of data from marketers to the DSPs of their media partners. This process facilitates the precise targeting of specific audience segments, ensuring that the right ads are delivered to the right consumers at the right time.
Brands and agencies can create more effective and personalised campaigns by suppressing or retargeting via onboarded data, while device graphs that associate device IDs with customer data enable a consistent experience across devices while minimising over-exposure through frequency capping. Onboarded data can also serve as a seed for lookalike modelling, which can generate predictive audiences that look and act like a brand’s existing customers.
While fragmentation and limitations in addressability mean the CTV space is still a work in progress, many platforms have developed capabilities for marketers to onboard their data directly (or via an onboarding partner) in a manner similar to a private exchange.
By integrating data from various sources, marketers can deliver more relevant and personalised ads to specific households. However, original equipment manufacturers (OEMs) and streaming providers are highly protective of their proprietary data and are unlikely to release it back to the brand or agency, which might consider third-party attribution partners instead.
Then there are the walled gardens, such as social platforms, Google’s advertising suite and — arguably — retail media networks. All allow marketers to directly utilise data onboarding to reach consumers within their closed digital ecosystems. As users typically sign up to such services via their primary email, match rates and accuracy tend to skew higher than the general digital environment. However, the limited data sharing within these ecosystems poses challenges for measuring campaign performance and attribution.
Wrapping it all together, anonymised data passed back to brands and agencies that onboard their data allows for in-depth analysis of cross-channel measurement and attribution, enabling them to evaluate the effectiveness of different targeting strategies and channels to make informed decisions for future campaigns. By collaborating with attribution partners, marketers gain a holistic view of the customer journey, understanding how different touchpoints contribute to conversions and overall campaign success.
Data onboarding was complex from the beginning and is no less so today as more and more channels and platforms enter the media ecosystem, while regulations and device-level privacy consistently upset established matching paradigms.
But brands today also have access to specialised onboarding partners and a growing number of data-savvy agencies that can help them with the heavy lifting. Whether they choose to manage it internally or externally, no brand can afford to ignore data onboarding in today’s omnichannel world.
Chief Growth Officer