As the holy grail of marketers since 2014, Data Management Platform (DMP) seem to have lost some of their shine. However, the issues they address are still just as relevant: reconciliation of offline and online data, audience enrichment and activation, personalisation of customer journeys, predictive model, etc.
It's perhaps because they got things right too early on that DMPs are now being challenged both at a legal level (GDPR) and in terms of the performance improvement they are supposed to deliver. Promises that were sometimes difficult to keep yesterday, such as cross-device, are now becoming a reality thanks to the technological leap made.
Even so, there are many challenges still to face. Although the introduction of the GDPR has so far had little impact on users'consent to cookies2, the arrival of the e-Privacy regulation could very well prompt a redefinition of how DMP players are positioned. This dynamic has already been initiated by the institutionalisation of Customer Data Platforms (CDP).
As a result, faced with a shifting regulatory and technological ecosystem, DMPs still have several benefits provided they can continue to meet customer needs.
The main function of a DMP is to collect, segment and qualify the audiences that browse on an advertiser's site. As such, scoring engines are natively integrated into most marketing solutions to enable the marketing teams to categorise audiences according to various criteria in order to activate them.
The more signals there are, the more relevant the segmentation can be. These profiles identified in the form of an anonymous ID are finally sent to various third-party solutions to be activated. Matching - very often in real time - based on a correlation table is then carried out, enabling the third-party solution to activate its own IDs.Empirically, there are many uses and these mean optimising the digital acquisition strategy:
One of the key challenges of these various use cases lies in the audience onboarding rate. Although DMPs can now ?cookify? profiles via three uses (open email ? registration ? login), the fact remains that a sizeable proportion of the audience does not engage in any of these journeys, meaning it might be necessary to use a CRM onboarding solution. The goal is very much to achieve a critical mass of actionable data.
To widen its prospecting universe, an advertiser might want to reach unknown audiences. To achieve this, it's in its best interests to use customer knowledge to target profiles very similar to its customers rather than the "run of the mill". In this respect, most DMPs allow advertisers to manage statistical twin audience segments based on a master audience.Once this lookalike segment has been generated, it can be sent to DSP third-party tools to conduct a media activation campaign. The benefit is twofold:
Notons également que la génération de ces segments de profils jumeaux peut se faire directement au niveau des DSP, dès réception de l'audience master.
The DMP "cookification" process enables more and more audience to be onboarded. The goal is to ultimately have the largest audience coverage rate possible, mapped in the DMP solution, so that personas can be built based on the behavioural history. As such, when a person returns to the advertiser's site, the information held will straddle both the CRM profile and the browsing behaviour. This amounts to relevant customer knowledge in which social demographic data lives side by side with transactional data.
As such, when a person returns to the advertiser's site, the information held will straddle both the CRM profile and the browsing behaviour. This amounts to relevant customer knowledge in which social demographic data lives side by side with transactional data.Starting from here, the advertiser can carry out scoring to identify types of engagement with which to associate personalised customer journeys. This means the user experience will be based on a 1:1 model, both at a site level and in terms of the marketing reach via third-party tools (DSP, marketing automation, etc.):
Traditionally positioned on the media side, DMPs have the fundamental advantage of interfacing with a multitude of data providers. Although some solutions are still very first-party data oriented, others give advertisers a marketplace that provides them with access to additional data segments. The key challenge behind all this is to build the most qualified and relevant audience segments in order to activate them through the most suitable offering.
As such, an advertiser may find it best to add an informal sociodemographic overlay to its site audience in order to most effectively personalise the user experience. Conversely, adding a knowledge layer to the identified customer engagement, through a record of journeys collected on other sites, is extremely valuable in enabling an advertiser to prioritise its advertising budgets.
Integrating third-party data into DMP segments is an expensive affair and whether this is relevant is still subject to the available reach. The recent introduction of the GDPR, as well as the upcoming implementation of e-Privacy, nevertheless raises questions about the volume of third-party data available. This data will probably soon be correlated with the user's motivated consent.
This is one of the reasons why second-party data deals between two advertisers are beginning to take place on a mass scale. Relevant from a business viewpoint and compliant with the GDPR, these foster extensive transparency in a complex ecosystem. In this respect, DMPs can serve as trusted third parties.
Often forgotten in the DMP equation, publishers nevertheless often face issues that directly involve data collection, segmentation and reconciliation. It's no coincidence that nearly 66% now say they have a DMP implemented in their ecosystem. According to the latest 2018 Programmatic Barometer by the EBG and Quantcast, 40% of publishers who don't have one have plan to do so.
The key challenge behind this technological maturation is to provide a more robust type of data, including, in order of priority, browsing, geolocation, demographics, CRM, third-party data, etc.
Faced with walled-garden ecosystems like Google and Facebook, which manage to capitalise both on declarative data and intent or behavioural data, publishers may find their salvation in DMPs. Once contextual, inventory monetisation offerings are increasingly being positioned from a customer-centric viewpoint. This is a shift from media planning to audience planning.
In fact, for a publisher, return on investment is judged based on the increase in CPM in its inventories. By providing increasingly qualified audiences, albeit smaller, the goal for these players is to meet advertisers' needs for 1:1 communication. As such, every web user will be offered, on a single site, a set of marketing touches personalised based on their persona. This is a windfall for publishers, which since the rise of programmatic have had to reduce their margin due to comprehensive competition.
At the same time, the implementation of DMP stacks is a way to reassure advertisers about the quality of the inventories available in programmatic. Indeed, advertisers' two main criteria are a desire to target the right audience and the visibility rate.
This isn't surprising, given that less than one in two impressions sold in programmatic are actually seen.Therefore, by securing a framework of high-quality distribution to selected audiences and the possible integration of a brand safety tool, a DMP provides reassurance to media publishers.
This introduction is an extract from the chapter dedicated to Data Management Platforms - Full Stack Focus in the Yearbook 2019 (Waisso) ; click below to download the long version: