Business data

Data

The word “data” is sometimes used in terms of a collection of data values. Data values can be “12-12-1967”, “John”, “0787654321”, “Amsterdam”, etc. In case of an electronic device, the volume of data values can be measured in terms of Kilobyte, Megabyte, Gigabyte, Terabyte, etc. So when we speak of “Big Data”, we mean high volumes of data values.

The word “data” is also used to refer to the type of data: “Order number”, “Order date”, etc. So when people ask “what data do you need”, they may expect a list of these data types. Within Business Data Management, with the focus on Business, we use the term “Business characteristic” instead of “data type”. Because we primary study the business world (e.g. “country name” when studying the globe and addressing countries) and not the data in accounting systems. Business characteristics are usually referred to by their name so it is important that these names are unique and homonyms and synonyms are avoided. So that business people really understand each other well for the sake of good semantics.

A third meaning of data can be on specific objects, like a single customer, an order, etc. So when people ask “what data do you need”, they may expect a list of these objects, usually asking for all data values available for these objects. Within Business Data Management, we also study objects but again: the real world ones, for example the customer we know by the name “John”. When we discover that a group of people (in general: a group of objects) behave the same, for example according to the definition of “customer”, we call “Customer” a Business concept.

Metadata

Within Business Data Management, instead of investigating data and creating metadata, we investigate the real world and create the Business Data Model from that. This model is descriptive, it is all documentation on the real world, it’s all meta. All information we collect on Business terms, concepts, characteristics, etc. we call “metadata” because it is descriptive data on the business aspects we study and model.

Before implementing Business Data Management in an organisation, one should think of the metadata required. It starts of with a name and definition for every aspect but many other descriptive metadata may be considered relevant. It is advised to create a Business Data Model on the metadata needed (which may be a good exercise on the metadata needed).

Remark: one could argue that there could be “meta” on data value level. We will not discuss this subject here.

(Meta) data ownership

When we speak of the owner of data, we mean the executive who is responsible for (the quality of) data values. Often a shortcut is made by making this executive also the owner of the metadata. In other words, by assigning a data owner to the data type (or: “Business characteristic” in BDM terms), this executive is owner of the metadata of the data type ánd of all data values assigned to this data type. Two remarks can be made on this practise:

  • There can be multiple data owners on the data values of a single data type. For example in case of multiple back offices on different locations (countries) where identical products are sold and administrated for.
  • The owner of the metadata (that is: on the Business characteristic) can be someone different from the owner of the data values.
    • For example a Product Manager defines the product, the back office administrating all contracts for these products, will have to comply to the Product characteristics. The Product Manager will be the metadata owner, the back office will be the data (values) owner.
    • It is possible that both ownerships come together within a single executive. For example in case Sales & Marketing has the ownership of customer data (the metadata) while the data (values) are maintained within the same department. The executive of Sales & Marketing will be the metadata owner and also data value owner. The lack of segregation of functions may lead to improved data awareness in case there is strong meta management, if not there is a risk that the data quality will suffer from it.

Business ownership

From Business Data Management perspective, it is relevant to know the business parties who are managing the data (the data owners) and who will be the owner of the metadata. Because it is relevant to know where and how the data on the real world is accounted for and by who. When modelling the real business world, we describe the Business in what we call meta business. That is: metadata on the business. There are some roles to distinguish:

  • Owner of Business. The “Commercial Party” running the Business. For example a loan department within a bank. Or the Human Resources department serving management and employees within the organisation. The owner will be the highest management of the Business.
  • Owner of objects. Some responsibilities may be delegated by the owner of Business to parties within or outside (by contracts) that business party. Think of sales, repair, call centre, etc. Within the organisation it might be managed by KPI’s or SLA’s, external parties are often contracted.
  • Owner of Business concept. That may be the R&D department developing cars for a car manufacturer, or product management creating the product Term loan for a bank.

The Business may have delegated the ownership in accounting & data for their business activities and results to administrative departments. The Business will always remain accountable for their business processes and systems. Both worlds will help to create the Business Data Model and will give insights in differences in what we do as Business and what we account for in systems.

Business: creators versus consumers

Companies spend a lot time, effort, money to maintain their assets. And that makes sense because when you need them, you want them to be there in good shape so you can do your work in a pleasant, professional and efficient way. Have a look into the van of a professional plumber or electrician, where and how tools and materials are stored, and you will know. Nowadays, with growing and changing emphasis on accounting, standards and reporting, in order to comply to regulatory demands, awareness has grown that data is an important asset for the company. From Business Data Management point of view we add the metadata to the valuable assets. Without the metadata semantics on the business and the data, without speaking one language within the company, communication & information and governing the company will be problematic.

The past 40 years we have seen a shift in data modelling and interest in metadata. Initially there was a focus on the primary business, their processes and efficiency. Data models of these data creating businesses were made to develop accounting systems and data to be stored and managed. There were conceptual, logical, technical and physical data models and there were attempts to integrate the conceptual and logical ones into Corporate Data Models on company level. Lots of time and money was spend and often these efforts were subject to reductions and cost savings. Data management departments were erected and later on often minimalised or eliminated when the biggest problems were solved (for the moment).

A next relevant development was the need to exchange data between accounting systems. At first that was because systems needed information from other environments and physical distribution of data through interfaces had to solve that problem. Or object orientation and the use of services could and did. The awareness of the need for efficiency, a single source of truth, led to the concept of master data management and reference data management with an increased need for exchange of data. Delivery of metadata (in any form, initially by means of manuals) was essential to understand what was delivered and how to read and process the information from the other environment.

A third wave in interest in data modelling came from the development and exploitation of Data WareHouses, where data comes together from all different sources within and from outside the company. Again there was an urgent need for an integrated model on company level. Not just on a technical level but also in terms of the business meaning of the data stored. Models bought might be considered to be a solution for the storage, maintaining and delivery of data, it will be out of sync with the business language and the Business Data Models of the primary businesses.

The growing importance of creating reliable information from data (from different sources), created a fourth wave in interest in data modelling and data management but from the data consuming perspective. Consuming departments like Management Information, Finance, Risk, filled the gap and were creating data models for the data they needed, according to their terminology, creating definitions, setting Business rules and Data Quality rules in order to get the right raw materials for their information creating processes. In fact, they took the role of the data owners and metadata owners setting their own standards. Creating new problems for the data creators because every data consumer talks its own language which to a certain degree will be different from the language of the people in the field.

There is only one way forward and that is that the primary, data creating businesses take their responsibility and make serious work of Business Data Management, creating their Business Data Model. They are the owners, not the data consumers. At the same time, there is nothing wrong in consuming parties creating their own lexicon and data models to specify their data requirements to explain: “this is what we need”! Especially for the time that the Business Data Models are still under-developed. It is a time consuming, expensive way for setting requirements but it will provide additional context to the lexicon when Business terms used in reporting, are added. Also in case when these Business terms have to comply to the terminology of external parties (e.g. regulators), it will be great to have these external terms added to the lexicon. It connects the internal world of the company with the outside world and that is beneficial.

In short: Business Data Management is developed based on the principle that any Business Data Model is a representation of the primary, data creating Business itself. Business terms are the basis, they should be named, defined and modelled.