CDP & DMP: how multichannel digital marketing is revolutionizing customer relations

[SPECIAL FEATURE] Since the early 2000s, the amount of customer data available has multiplied, and with it the possibilities for marketing actions. However, traditional marketing solutions (CRM, email marketing, etc.) are far less relevant and unable to manage all this data in a unified way. DMPs and, a few years later, CDPs were born out of this fascinating problem. In addition to aggregating all your data, they promise to put it into action across all your channels. Without further ado, here's an overview of these two marketing trends that many experts see taking off in the coming years.

CONTENTS

See all software Customer Data Platform (CDP)

Digital strategy and the customer journey: from CRM to omnichannel marketing

Historically, the first tool to centralize customer intelligence was CRM (Customer Relationship Management) in the 90s. This was followed by direct marketing tools (email marketing, SMS marketing, content marketing) in the 2000s. It's only in the last 10 years that platforms exploiting online data (DMP, CDP) have developed, making it possible to take advantage of Big Data to improve customer knowledge. What customer data platforms (CDP) and online data management platforms (DMP) have in common is the promise of a complete view of the customer. We'll see below that for CRM and DMP, this promise is not entirely fulfilled, but that in the case of CDP, this card remains to be played:

Definitions

Definition of CRM or Customer Relationship Management

CRMs are customer relationship management tools that have been used in the direct sales and loyalty sectors since the 1990s. Their aim is to centralize customer knowledge around a CRM identifier in order to improve the quality of customer relations. In terms of marketing data, CRM provides a nominative view of customers and prospects.

Benefits of CRM

• CRM enables you to automate sales and loyalty actions with identified individuals, thus boosting productivity and competitiveness.
• Data has a higher unit value than in a CDP/DMP, because it is qualified by humans.
• CRM data can be used to initiate direct marketing actions rapidly (email campaigns, SMS).
• No technical skills (Business Intelligence, Data Mining) are required to exploit this data.

CRM drawbacks

• CRM is limited to offline data (does not aggregate cookie data, for example).
• CRM is a "single channel" tool only, unlike cross/multi/omni channel platforms.
• CRM is limited to structured data only (which represents a very small proportion of available customer data).

Which companies use CRM?

• All companies selling goods or services are concerned by the use of a CRM.

What does a CRM tell us about a customer?

• Name, gender, CSP, family background, age, interests, purchase frequency, etc.

Definition of DMP or Data Management Platform

The Data Management Platform (DMP) is an aggregator of anonymous data reconciled around the cookie. For simplicity's sake, we can think of it as a very large database. The DMP is designed to better operate digital campaigns (programmatic) in an open world. The DMP is based on Hadoop, Hbase, Map Reduce and other technologies, and offers a statistical view of visitors and audiences.

DMP benefits

• Precise audience targeting
• Integration of a DMP is simpler and less costly than a CDP.
• DMP allows you to separate yourself from advertising agencies, and gain direct access to precise data to manage the most relevant marketing actions (channel, banners).
• The cookies collected can come from both proprietary sites (first-party data) and affiliated sites (third-party data).
• Some DMPs have connectors with third-party tools such as CRM, marketing automation (like Marketo) and DSPs.
• The DMP enables us to understand the customer journey, so we can adjust RTB campaigns across different channels.
• It can also calculate the return on investment (ROI) of campaigns by channel
• It can manage raw, uninterpreted data (dates, receipts, GPS coordinates, etc.).

Disadvantages of DMP

• Offline data is missing to achieve the customer knowledge objective (non-media value).
• The data retention period is only a few months. This is too short to understand a customer's life cycle.
• Data is not cleansed (raw and sometimes duplicate data), making access and marketing activation more complex than in a CDP.
• There is little or no possibility of implementing direct marketing actions from a DMP.
• Data mining and machine learning functions are either non-existent or very rare.
• DMPs are only of interest when 100,000 or more cookies have been collected.
• It offers no real real-time data management.

Which companies use a DMP?

• Retail companies (ecommerce and physical stores), travel and tourism companies, affiliation agencies (for more detailed retargeting and campaign optimization), as well as banks and insurance companies.

What does a DMP tell us about an audience?

• gender, age, interests of content consumers

Definition of CDP or Customer Data Platform

The Customer Data Platform (CDP) is a solution for aggregating and putting into action all online and offline customer data. The CDP, born in 2013 from David Raad's concept, makes it possible to create an integral knowledge of the customer. Like the DMP, the CDP is most powerful in programmatic (media), but can also be used to activate other marketing levers. CDP provides an individual view of customers, prospects and visitors.

Advantages of CDP

• Optimize multi-channel campaigns and reach.
• Allows marketing professionals to easily manipulate data (vs. business intelligence professionals for DMPs).
• It provides a 360° view of the customer, and offers highly precise targeting leverage.
• CDP can be used to create precise segments and analyze purchasing behavior.
• It enables precise multi-channel ROI analysis and cross-channel repercussions.
• It facilitates compliance with the European regulation on personal data management (RGPD) mandatory from May 2018: centralization of personal data, data delivery on request and data purging.

Disadvantages of CDP

• Complex integration: setting up a CDP mobilizes a large part of the company for 3 to 6 months (business lines, Information Systems Department, managers, legal department).

Which companies use CDP?

• Companies with siloed data, mature data processing and substantial financial resources (e.g. SNCF, SAMSUNG, etc.). These companies will make several thousand segments per year.

What does a CDP tell us about a customer?

• Time spent on a web page by an identified or unidentified user, email open and click rates, interest in a subject, relationships, etc.

What are the differences between a CRM, a CDP and a DMP?

We've seen the major structural differences between CRM, DMP and CDP, as well as their advantages and disadvantages. Here's a functional comparison of the 3 platforms:

*ROPO: Research Online, Purchase Offline

MarTech, AdTech: welcome to the era of Data Marketing

Why use a CDP or DMP?

According to the Marketing Agency Growth Report (Hubspot, 2018), 37% of agencies are unable to attract the ideal customer, and 39% of these agencies don't abandon a business relationship even if the prospect's profile doesn't fit their model. This is exactly the problem that CDP and DMP solve.

CDPs and DMPs are aligned with the abundance of data and the opportunities and constraints inherent in this context. Let's be more concrete: here are 6 reasons to choose one or the other of these two solutions.

1. Multiplication of channels

All consumer activities have a digital footprint, even the most mundane. The simple act of shopping generates loyalty card data that is then widely exploited. There are hundreds of different channels, all of which necessarily have an impact on each other, and the only way to collect and unify this data around unique profiles is via a CDP or DMP platform. For example, they can create synergies between channels by targeting newsletter subscribers who don't open emails with banner ads (thus breaking down the silos between ad buying and emailing).

2. Segmentation

The relevance of a customer, prospect, visitor or audience segment is only really relevant if there is sufficient data. For example, segmenting white shoe buyers is not really relevant. However, segmenting buyers of white shoes who are about to get married is much more relevant. On the media side, DMPs help optimize investments by excluding customers who have already purchased or are overexposed. In short, CDPs/DMPs are the only tools that can be used to set up ROI-effective segments, thanks to the sheer volume of data and the richness of their profiles.

3. Activation

Data aggregation platforms can transmit segments to marketing automation or programmatic tools, or even activate marketing actions themselves. This centralized orchestration of marketing actions is the only way forward for marketing. On the one hand, because it is becoming standard practice in all commercial enterprises, and on the other, because irrelevant marketing approaches are very poorly accepted by consumers (spamming, complaints on social networks, depreciation of brand image, etc.). Consumers have become accustomed to individualized marketing: this is the norm.

4. Tools for marketers

No marketing professional accepts having a partial, blurred and therefore almost certainly erroneous vision of an audience or customers. It's impossible to commit to marketing performance when the information system doesn't feed marketing with complete, honest data, just as it's difficult to be a good cabinetmaker in 2018 without laser cutting...

5. Personal data management

No hosting provider today agrees to host personal data anymore, and for good reason: data leak scandals (including Facebook's on March 16, 2018) are repeating themselves and burning the hands of those who must ensure their protection. What's more, data protection legislation is tightening up everywhere, the most important of which is the RGPD (General Data Protection Regulation). To be compliant, we need to be able to control the completeness of personal data (reconciliation, transmission, deletion) which is the basis of a CDP.

6. Behavioral prediction

Data collection is the only way, thanks to artificial intelligence or data mining, to detect purchasing intentions (thanks in particular to browsing data).

Which customer knowledge management platform to choose?

Let's not make the mistake of choosing a solution by looking at its data; it's quite simply the biggest mistake possible. The choice between DMP and CDP depends on whether you want to activate your data online or offline. And although there is a functional overlap between the two platforms, DMP clearly tends towards online activation, while CDP tends towards offline activation.

In this logic, you don't necessarily have to be desperate for a holistic view of the customer (a kind of holy grail sought by the most technical among us). On the other hand, the relevant personalization of the touch point is a sound objective and an excellent starting point for drawing up project specifications.

This corporate objective, which is also the starting point for the project, must be defined by the business line, which must also be integrated into the entire DMP/CDP lifecycle (data selection, data quality, segment creation, data activation, etc.). All other specialties must also join the project on an ad hoc basis (legal, IT, management, etc.).

What impact does Data Science have on marketing managers?

Marketing's role in the customer acquisition and loyalty cycle has steadily increased. Just a few years ago, marketing focused on introducing consumers to a product and stimulating their interest. The rest of the cycle was taken care of by "salespeople". Today, webmarketing knows and must also stimulate consideration, purchase intent and product trial. Among its KPIs are conversion rates and campaign profitability.

This evolution is due to the increase in the mass of data and the tools for exploiting digital footprints. As a result, the marketer's profile is evolving: from a creative profile looking for ideas for marketing actions, to a decision-maker who makes choices based on the information he sees (falling ROI, growth of a channel, repercussions of a campaign from one channel on another, etc.). In 2020, the ideal marketer will have a scientific approach to marketing, with the empathy to make the right decisions. Data science will be a key skill for understanding the predictive algorithms that will recommend marketing actions.

Is this already the end of DMPs?

Criticism of the model

Before we talk about the end, let's talk about the beginning of DMP. The promise of the DMP was to bring together all customer data, when in reality this was never the case. DMPs are platforms for aggregating contact data of low unit value only. This is great enough for all the companies that use the media to promote their products (and there are many of them), but this vagueness has been the cause of many, many disappointments. Some companies who thought they could access all their customer data at the snap of a finger with DMPs have met with painful failure.

The notion of easy access to data brings us to the second reason why some DMP projects fail: the complexity of accessing data. As we saw in the table above, the DMP is fed with raw data that can be exploited by technicians (Business Intelligence Analysts, Data Scientists) who are also the ones who set up the DMP. Yet the success of a DMP depends on the resulting marketing success: it's the marketers themselves who must have easy access to all the data they need, which is unfortunately rarely the case.

The third reason for DMP setbacks is the underestimation of the effort involved, including the difficulty of unifying data, the time it takes to train teams, the time it takes to set up, and the costs. These are just some of the reasons why DMP projects are so frustrating.

Last but not least, many DMP projects have never been profitable, as there is a minimum threshold of data required to operate the platform properly. For some, however, the volume of media purchases, and in particular retargeting (third-party), is not sufficient to make their DMP profitable.

Double-digit growth by 2020

Nevertheless, we won't be seeing the end of Data Management Platforms any time soon. The above comparison table shows that CDPs are not able to handle as much data as DMPs, and that access to raw data at low unit values is still useful for understanding things. In addition, DMPs have the great advantage of making advertising budgets transparent (spending, ROI, campaign management), as was previously the case with agencies. Perhaps we should be talking about Audience Management Platforms rather than DMPs.

The first part of this video explains the subject of this article very well.

Between DMP and CDP, which marketing solution should I choose?

The market for centralized marketing data (Data Marketing) is highly protean, as all players - CRM, Marketing Automation, DMP, CDP - want to offer this holistic vision of the customer. This is the case, for example, with Salesforce, which covers all channels with Sales Cloud for CRM, Service Cloud for customer service, and Marketing Cloud for SMS, email and more. Marketing giants are also embarking on this quest, such as Marketo with Turn (its DMP) and HubSpot with its increased functional coverage. Here are the main DMP/CDPs on the market:

• Adobe Audience Manager ;
• Cxense ;
• KBM Group (Zipline DMP) ;
• Krux ;
• Lotame ;
• Neustar (PlatformOne) ;
• Oracle DMP ;
• Makazi ;
• Weborama ;
• Ysance ;
• Mapp Digital ;
• Eulerian Technologies ;
• 1000Merci ;
• Quintessence (Base Camp) ;
• Cabestan ;

Each of these players has one or more specialties that you need to look at carefully to make the right choice. Camp de bases, for example, offers a high standard of Data Quality and supports its customers with a proven method (Data Deep Dive) that enables them to rapidly achieve ambitious objectives. Please refer to the list of DMPs and CDPs on appvizer to compare the different solutions.

See all software Customer Data Platform (CDP)

Conclusion

We have seen in this article that CDP and DMP solutions make it possible to unify marketing data and activate it across multiple channels. This centralization of data management and marketing actions makes it possible to create smarter campaigns offering better returns on investment. Competition between DMP and CDP is often evoked, but we have seen that, despite a functional overlap, the former's purpose is online activation (media), while the latter's purpose is offline activation (direct marketing).

Although data culture and data marketing solutions are relatively new, the market is now ready to accelerate, with a direct impact on marketing professions. Marketing professionals are becoming more decision-makers than creators.

By 2020, Data Management Platforms and Customer Data Platforms will evolve towards the integration of predictive models (Smart Data) and the unification of cookies (audience data) with CRM data (personal data).