Companies are starting to question traditional marketing paradigms, and explore the use of ultra-personalisation tools and methods.
However, while this approach is intellectually stimulating, it is not without its technical and conceptual pitfalls. The lack of maturity of DMP media in the market, the limited automation of certain creative aspects of campaigns, and the perspective required to intelligently configure messages in complex and congested channels.
The concept of ultra-personalisation can be hard to grasp as its very definition and range of possibilities varies depending on the advertisers’ business activities. For example, a cosmetics brand would not employ the same ultra-personalisation strategy as an airline or rail company, even though they share the same audience! Stores, stations and airports are all points of contact where transactional data and tour operators’ data could be collected to define a targeted ultra-personalisation strategy. Unfortunately, ‘second-party data’ use cases are rare and still in their infancy, so it is necessary to employ other approaches to ensure relevancy for customers and prospects.
Data quality and access to sensitive and personal data also presents significant challenges. The ultra-personalisation approach can pose ethical problems, which will be further highlighted following the implementation of the GDPR in May 2018. Advertisers are becoming aware of the need to develop a more measured approach via permission marketing. The use of data science is no longer just a lever for individualisation but also a means of creating relevant campaigns based on more aggregated and less sensitive data.
Among popular data-driven approaches, mix modelling is a robust solution traditionally used to explain performance drivers. New and advanced modelling approaches make it possible to identify the right audience and move from target planning towards audience planning. By understanding customer behaviours and ensuring strategic ROI management, audience planning enhanced by modelling can be used to break down the ROI e.g. by customer segment, purchasing funnel level, and type of product. One of the most significant benefits of this approach is that it can be applied to all activation levers rather than solely ‘one-to-one’ levers.
However, exclusively focusing on an ultra-personalisation approach may have led to marketing myopia, as well as having a significant impact on the profitability of implemented actions due to an overly ‘micro’ approach and one that was affected by the quality of the data and the execution options. Therefore, it is necessary to take a step back to understand how data science can add relevancy at all decision-making levels, rather than seeing it as generalised and utopian ‘one-to-one’ research.