For many companies, the reality of building management information reports and performing business analytics is similar to conducting your weekly shop from a traditional town centre. It is time-consuming visiting multiple shops and paying at each in turn. Similarly, data sitting on disparate systems forces teams to manually stitch and manipulate datasets into usable information from various locations.

Product lines between shops invariably overlap, causing shoppers to make multiple choices on the best place to buy each item. Likewise, data duplicated across a companies ecosystem makes reporting and analytics self-service challenging.

A far more sensible approach is having all your data in one place for reporting and business intelligence activities. But traditional data management approaches may struggle to resolve information siloing. It is inherently tricky to combine data from disparate systems correctly. So you may simply decide to drop your data into one area without integrating it in any way. Looking at our shopping metaphor, this is equivalent to placing a large roof over the shops. They are still just a collection of individual shops under one roof. You haven’t resolved the underlying issue.

A supermarket makes shopping far easier. Vast product lines which may have originated from multiple suppliers are logically grouped within aisles, making it quicker and simpler to locate the items you require. Modern data management methodologies work similarly. Data from various disparate systems can be integrated around business concepts, such as customer, store and product, to produce meaningful information.

To logically layout a supermarkets’ aisles, carry out product grouping and placement takes careful planning. Get it wrong, and the shopping experience can be confusing at best. At worst you may choose to shop elsewhere. An enterprise data model is used to represent “a single integrated definition of a company’s data” (Noreen Kendle, 2005), by uniting the business concepts important to an organisation. Designing a data management architecture around this model ensures data is grouped and ordered in a way the business understands, making activities like self-service much less stressful.

You may prefer to order you’re shopping online and have the supermarket deliver it to your door, further freeing up your day. The supermarket can achieve this because it has logically grouped the thousands of product lines. A store picker can grab items across multiple shopping lists with ease. Having a sound data management architecture in place makes it’s possible to automate data activities. Machine learning can identify trends and patterns in data or assist with continuous improvement initiatives, for instance.

Taking the time to get your data in order, will pay dividends down the line.