Click to learn more about author Stuart Tarmy.

Data Flow allows a company to understand the relationships between data elements across the enterprise and between individual lines of business.

Companies in a wide range of industries, including financial services, health care, aviation and energy services, are increasingly focused on Data Flow to better leverage their Data Analytics to improve cross-selling, and to ultimately differentiate from their competition. If your company is not conversant, actively engaged, or soon to be actively engaged in Data Flow, then you could fall behind in a key information technology wave.

What is Data Flow?

Data Flow, sometimes referred to as Data Lineage, describes how data moves, or ‘flows’ across and through an enterprise. It can be at a very granular level, such as providing an audit trail that literally shows how a data element (like someone’s last name) interacts with other data elements before reaching its final destination. Or, Data Flow can operate at a high level, showing how a data element interacts with other systems or databases within an enterprise.

Similarly, users can understand Data Flow at either a granular or high level.  At the granular level, usually the concern of data scientists and data architects, Data Flow can be represented within a database by the tables, schemas, columns, data elements and the primary or foreign keys that link them.  At a high level, usually the concern of business users and analysts, Data Flow can be represented visually, for example using different size circles showing the data tables, schemas, columns and elements and using lines or arrows to show the connections between them.

Why Data Flow is Important?

Data Flow is enormously important for analytics and cross-selling because it allows a company to understand the relationships between data elements across the enterprise and between individual lines of business. Understanding Data Flow can help answer questions such as: From where did the data originate? When was it recorded? What systems are using the data? and What business units are using the data? Being able to answer these questions and to share information across an enterprise provides enormous benefits for analytics and cross-selling, such as enabling a ‘360 customer view’.

Unfortunately, many business units, functions or departments within a company are often run as independent silos, only able to ingest, process and report on data within their own four walls. The technology architecture does not allow them to easily share information with their other business units, creating severe competitive disadvantages for the company and forcing employees to sometimes implement awkward, inefficient and costly workarounds.

Data Flow can Enable Analytics and Cross Selling

Cross-selling is defined as encouraging a customer who buys a product (e.g. a large screen TV) to buy a related product (e.g. an audio sound system). The obvious benefit to cross-selling is increasing revenues at a lower customer acquisition cost because the customer is already familiar with the brand. Data Flow is a prerequisite for understanding the data relationships, and hence the ‘360 customer view’, across an enterprise.

Imagine a financial services firm has multiple business units that are siloed, such as a retail bank, commercial bank, brokerage, high net worth group, credit card group and a mortgage loan unit. While each of these business units may have a good understanding of their own siloed data, the company at the corporate level is not able to generate a ‘360 degree’ view of their customers.  Because of this, they can’t identify attractive cross-sell opportunities, such as an existing customer who has a large sum deposited in his brokerage account who would be a great prospect for their high net worth group. Or the bank customer who has taken out multiple real estate mortgage loans who could be referred to their commercial bank.

In addition to increasing revenue, a less obvious benefit to cross-selling, but equally powerful, can be in customer retention. For example, according to a Bank Intelligence Solutions (BIS) study from Fiserv, a customer who uses one bank product will remain a bank customer for about 18 months. But, if the customer buys an additional product, the customer relationship is increased by about four years. If the customer can be persuaded to purchase three products, the average banking relationship increases to almost seven years.

Machine Learning-based Data Discovery can Break Down Silos to Enable Data Flow

For companies with siloed businesses, there are several solutions that can break down these silos. The oldest and most well-known solution is a manual approach, whereby a company will assign staff or hire outside consultants to manually research all of the company’s databases. This manual process can be tedious, costly and error prone. Our research has shown that a good database expert can sometimes only review approximately 2-3 databases per month on average. For larger companies that may have thousands of source databases, this is not a realistic solution.

Alternatively, innovative technology companies have addressed this problem by pioneering machine learning driven smart data discovery solutions that can automate the Data Flow discovery process across large enterprises. This data discovery process works directly with the data itself, and is able to understand the Data Flows across all of the company’s systems, including the more problematic siloed and legacy systems.

Other posts by Stuart include:
To Get Value from Data, Data Discovery Must Come First!
Appoint a Data Protection Officer to Ensure Compliance with the GDPR