Comment on page
Welcome!
Welcome to our documentation web site for a technical audience.
Whether you are Ad / Digital Ops, Data Engineer, Data Analyst or Data Scientist, you should find here all the information you need to make the most of the mediarithmics platform.
Our vision is to provide a data-first marketing cloud to solve the need for organisations to serve as best-in-class in a digital economy.
The mediarithmics platform can be seen both as a set of digital marketing applications that can be used directly by marketers, or as a highly customizable data marketing cloud infrastructure that data engineers and data scientists can tailor to the needs of their organisation.
Even if this 'Get started' section is mainly targeted at a technical audience, it is written in plain English, without reference to technical implementation details, and should be readable by anyone interested in the way mediarithmics structures the world of modern marketing. It is a good introduction to the mediarithmics platform.
The purpose of a data marketing platform is to enable use of all available information to make better decisions in all areas of marketing.
The mediarithmics platform can collect all data sources. Whether the data comes from the users' online activity or from backoffice management systems such as CMS', ERPs, and CRMs, all the data is stored for each user in a graph model as shown below. This universal data model allows you to connect all the information of a user, the collected information such as account identifiers, terminal identifiers, profile information, activities, and purchases, as well as data calculated on the fly, such as appetence score, age and gender predictions, and recommendations of articles or other content.
These centralized data allow us to conduct multiple analyzes (visit analytics, uplift analytics, segment insights, data discovery ...) and to act when the time comes:
- Either by activating audience segments via a rich ecosystem of connectors to advertising networks, communication tools (notifications, emails) or personalization tools (on-site display, A/B testing on user journey),
- Either by defining orchestrations that will be triggered automatically, by aligning themselves with the highlights of the user experience (e.g., first visit, creation of an account, drop in the frequency of visits, etc.).
In many companies, the recent history of data marketing has been built by the accumulation of different solutions that stack or overlap. Whether it is for the acquisition of new customers, the improvement of the conversion rate, the retention of existing customers, or the construction of predictive models, each application in isolation makes sense and provides a service to a team.
But after a while, this structuring in silos no longer makes it possible to be reactive and to innovate. The data is multiplied in multiple systems and these systems cannot speak to each other because of the "heaviness" of these huge amounts of data.
The mediarithmics platform has been designed from the start to overcome these structural difficulties and introduce a new model: the data-first marketing cloud.
In this model, the data gathered in a single datamart are shared by all applications and users. Both mediarithmics and partner applications access the same source of truth for analytical queries that mobilize large amounts of data, but also for all real-time personalisation micro-queries that are interested in the characteristics of a single user.
All the user data captured or calculated by the mediarithmics platform ends up in a special database named Datamart.
A mediarithmics datamart is more than a traditional database. This new generation database is leveraging a database engine specially designed to outperform in the domain of data-driven marketing applications. This database engine is the outcome of more than seven years of R&D from mediarithmics engineering teams. It is the first database engine to provide both real-time query processing and scalability on large data volumes (see focus on Datamart).
The mediarithmics datamart is a key enabler of all the use cases that can be imagined in one-to-one marketing … (real-time user experience customization, audience segmentation, funnel analytics, and campaign analytics …).
The information inserted in a datamart is primarily stored following a graph structure, where nodes containing pieces of information are connected through a link to represent a relationship between those pieces of information.
Let's take an example. A user visits a website which implements a tracker connected to the mediarithmics platform. At the end of the visit, three nodes will be created in the datamart.
- One node to represent the user
- One node to represent the visit (e.g., piece of information: the web site, the date, the visit duration, pageviews, products in the basket, etc.)
- One node to represent the device which was used.
Notice that the node representing the user looks like a pinpoint. This node is called a User Point and it plays the role of a central pinpoint connecting all pieces of information which are collected about the user.
In the User 360 View, it is important to distinguish the nodes representing a user identifier and the nodes representing a describing content on the user.
- The User Account Id : It represents the identifier of a registered user. The registration system can be a CRM system, a loyalty program system, or an authentication system. This registration system is external to the mediarithmics platform. The User Account Id is composed of a character string and a compartment id, which represents the registration system.
- The User Email: It represents an identifier based on the user email. It is usually derived from the user email by using a hash function (MD5, SHA-256, …)
- The User Agent: it represents the identifier of a user device.
The describing content of a user 360 view is composed of different data elements corresponding to different point of view on the user:
- A user activity represents an interaction with the user
- A user profile represents a summary of timeless information about the user (e.g. firstname, lastname, birthdate, address, sex, age, etc.)
- A user segment is here to capture that the user belongs to a particular group of users, such as an audience segment.
- A user choice is here to capture that the user has given his consent for a specific type of data processing
- A user trait represents a calculated attribute used to describe the user
A User Activity represents any interaction with the user either online or offline. It can represent the summary of an online visit, the detailed content of an in-store purchase, the summary of a call to a call-center, a campaign interaction like an opened email or a click on a banner.
A User Activity can contain a list of events. An event represents the most granular level of interaction with a user. It is defined by a name and a set of custom properties.
For each user, the user activities are ordered in a timeline.
For more detailed information on user activities and user events, please follow the link below.
User Profile
A User Profile provides a summary of information on the user and can contain any kind of information. The user profile is usually imported from an existing information system like a CRM or login database, and collect information such as: contact details, the user preferences, the status in a loyalty program, the subscription to newsletters.
A User Segment node represents a user who belongs to an audience segment. An Audience Segment is a group of users who share some common characteristics in the eyes of the marketer. There are several ways to define an audience segment in the platform.
Whatever the type of Audience Segment, a User 360 view contains a User Segment node for each audience segment the user belongs to.
For more detailed information on audience segments and their life-cycles, please follow the link below.
A User Choice node represents the choice of a user regarding data privacy and data processing declared by the organisation. This information is used to automatically adapt the behavior of the applications to consider the user choices about Data Privacy.
A User Score is a set of data calculated dynamically and providing predictive information on a user. A User Score is connected to a Machine Learning Function (ML Function). A User Score can, for example, predict the expected lifetime value, churn risk, age, or gender of a user.
All platform services are directly manageable and consumable through APIs.
Beyond these standard capabilities, the platform is also a computing infrastructure which enables the hosting of custom algorithms packaged in plugins. A plugin is a bundle of code and data which defines the behavior of a live function hosted in the mediarithmics cloud. It can be developed in any programming language. Due to latency requirements, some languages may be more adaptable than others.
The mediarithmics platform can integrate these plugins:
- Activity Analyzer: to filter and enrich activities data collected through a tag or the API
- Audience Feed: to push audience segments to third-party systems
- Machine Learning Function: to add live predictive data to a user graph
- Display Ad Renderer: to customize the rendering of display ads
- Email Renderer: to customize the rendering of emails
- Attribution Processor: to customize the attribution of a marketing conversion to different campaigns
The mediarithmics platform operates at the heart of an open ecosystem by offering connections with many partners.
Our platform covers your needs, including (but not limited to): analytics tools, messaging solutions, personalization solutions, prediction tools, advertising networks, and audience monetization.
Last modified 4mo ago