Unlocking the Enterprise Data Landscape

Unlocking the Enterprise Data Landscape

August 2019 | Marketing Technology

Data is complicated, and big data has made it more so. A vast array of databases exists to cut through the noise. Make sure to use the right one for the job.

Any discussion of integrations and data management inevitably gets expanded to include references to a central data storage location controlled either by IT or a dedicated marketing data team. Terms typically used to refer to this system can include data warehouse, a data lake or MDM. Data Warehouses and Data Lakes are the embodiment of Big Data, the hottest enterprise trend of 2014. They are intended as a repository to collect all data possessed by the business in one place so that it can be analysed by data analysts and data scientists to unlock trends and patterns that can improve marketing results and enhance corporate performance. As such, they can be considered the 'data' part in the decade old trend of data-driven marketing.

To the typical marketer, these tools are a mysterious black box containing vast quantities of data that could be used for campaign segmentation or analytics if the high guardians of IT weren't on hand to stop them. Those access restrictions typically exist for a reason. Data is complicated, particularly at the terabyte scale of a global business earning millions of dollars a quarter. It takes a highly trained specialist months to understand all the data sources within an enterprise and what they can be used for. The average field marketer would be quickly overwhelmed by the array of information contained in the typical data warehouse.

More importantly, data protection laws limit how the contents of a data warehouse or data lake can be used. GDPR was drafted to explicitly prevent organisations from using data collected for operational reasons for other uses, meaning that data in the data warehouse such as product health monitoring or usage analytics can't be used for marketing. Data warehouses and data lakes contain the full scope of the enterprise data landscape, so will include data sources that legally cannot be accessed by marketers. To prevent security breaches and compliance fines, access to these critical data storage locations are heavily locked down to the data analysts and IT staffers who absolutely must have access. Everyone else is given a locked-down feed or a BI dashboard with the information they need to know to do their jobs.

The exact level of access will vary by type of tool. Data architecture is a complicated business of critical importance to the modern enterprise. Enterprise architects are highly sought after for their ability to harness the complex web of technology and data in ways that can empower the business. There are many different types of database, each with their own core purpose. Most companies will have more than one, and it is essential to understand the difference in order to make the most of them.

Data Warehouse

Often abbreviated to 'EDW', enterprise data warehouses have been around for decades. They were pioneered by IBM at the earliest stages of the modern computing revolution and have been a corporate mainstay ever since. They were intended as a mechanism for organising all the data held by an enterprise into a highly organised, normalised structure so that it can be used by the business for analysis and decision-making. Usage of data warehouses has been on the decline, with specialist tools taking over many use cases.

Data Lake

One of the newer data storage technologies, data lakes are a product of the cloud computing boom. The most widely used data lakes such as Azure Data Lake, Amazon S3 or Hadoop are cloud technologies. Like the data warehouses that proceeded them, they are designed to act as a central repository of all enterprise data so that it can be used for analysis or decision-making. The difference between data warehouses and data lakes is that the data in data lakes is unstructured. Any data loaded into a data lake is left in its raw format, and no attempt is made to organise that data or clean it.

Master Data Management

Commonly abbreviated to MDM, Master Data Management is both a methodology and a technology. MDM is a recent offshoot of data warehousing, intended to make the data contained in a data warehouse more useful for the business users that need access to it. They are a reaction to the fact that most enterprises have too much data for the average business user to understand. Master Data Management is the process of taking the most important information out of all of the various data sources in the enterprise and organising it into a single easily understandable database that can be used by the entire business across all functions. This can include finance and operations as well as sales and marketing.

Data Management Platform

Commonly abbreviated to DMP, Data Management Platforms are most frequently used in B2C marketing and B2B advertising. They are used to build audiences for digital advertising campaigns, which are then synced to DSPs when the time comes to launch a campaign. Unlike MDM systems, DMPs typically only deal with anonymous web visitors. Most DMPs are simply pools of third-party advertising cookies dropped by the platform vendor on the devices of web visitors. The DMP vendor then builds a profile of the individual in front of the cookie based on IP information and the data collected by the cookie. Only rarely is the data in a DMP integrated into the broader enterprise technology landscape. Few DMPs attempt to link the data they contain to known contacts or customers in the enterprise as the cookies behind the DMP are controlled by the vendor rather than the customer, and in some cases may be used by all of the DMP's customers.

Customer Data Platforms

The newest type of data platform in the B2B marketing tech stack, CDPs are marketing-centric databases designed to collect all the information that the marketing department holds about all customers and prospects in one central database. Unlike the other tools on this list, CDPs natively handle account, contact and anonymous visitor information and aren't supposed to privilege one data source over the others. They build a single customer view, including a harmonised of view of both profile data and activity data across the entire marketing technology stack. This is then cleansed and normalised using AI so that it can be used for segmentation across all marketing channels. The leading CMS vendors have had CDP integrations for a while, and now the leading marketing automation platforms are following suit with both Oracle and expected to add integrations between their marketing automation and CDP products in the next six months. The idea is that CDPs will bring the different marketing channels closer together, and link them all together with the rest of the technology stack in a way that the other technologies on this list haven't been able to. It's still early days for CDP adoption in B2B marketing, but the technology has seen great success for B2C marketing.

Written by
Marketing Operations Consultant and Solutions Architect at CRMT Digital specialising in marketing technology architecture. Advisor on marketing effectiveness and martech optimisation.