Data Value Chain – Reference Model
My previous blog looked at datafication, a recent trend that suggests treating data as an enterprise asset. The blog proposed some extensions to the DAMA functional model.
This blog presents a new Data Value Chain- Reference Model. As stated before data is to be treated as an enterprise asset. And as assets represent a value we would like to relate the value of assets to each other in a visual way. The model displays common data sets / stores and data related functions in sequence of increasing value. It proofs to be a useful communication-tool for positioning our services and take the customer through the complexity of the data-related landscape.
The model can be considered as a reference architecture, containing the most relevant data-sets, functions and flows from our point of view – considering technical developments in the market.
Below I will briefly discuss the model:
General
The model is an overview of the complete IT-capability landscape related to Data. It depicts the relative value of data-elements (functions, sets, stores) in increasing from: From left to right. Stages are used to horizontally divide the model. Each stage groups elements with similar value. Elements may also be grouped functionally by (common) concepts, like IoT or Analytics. Concepts may cross stage-borders.
Concept-groups or elements may be mutually linked by arrows, representing data-flows that are considered as common in this context.
Note that the model does not pretend to be exhaustive: It may miss certain less common elements and flows. (See the explanation below)
Stages
Stages group elements in order of increasing value (left to right). The model contains the following stages:
- External data sources: Data that is external to the enterprise, which needs to be acquired before it can be used.
- Data Production and Acquisition: In this stage data is either produced (manually or through systems), or acquired from external sources
- Basic Data: Data that has been gathered in the previous stage is stored here in a crude state to be used for any preparation or enhancement action
- Enhanced Data: Basic data has been enhanced (processed, cleansed, transformed, aggregated, enriched, ect) so it can be used for analytics or BI purposes
- Intelligence: Data is used to run analysis on it to create business intelligence and decisions
- Data Products: Data products have direct value for external data consumers and may be applied for commercial purposes
Some elements and concepts may cross stage-borders as they apply to both stages.
Concepts
Concepts are familiar IT related terms in the market that are used to group elements. The purpose is to organize the model properly and map IT capabilities more easily.
Elements
Elements may be either data related functions or data sets / stores. The model distinguishes both, however this is not always a very clear distinction (e.g. Content Management consists of both data and functionality). Separating them would make the model unnecessarily complex)
Note: elements may in fact occur multiple times, e.g. in a recursive way: E.g. an IoT platform hub may contain a complete big data / analytics solution. These details are left out in the overview but may be added when zooming into IoT.
Data flows
Data flows describe the common flows of data between elements or concepts. As this is a value-chain the direction is mostly downstream, indicting increased value generation. Note that in real-life for example data usually flows back from MDM to Enterprise Systems. For readability only common flows are displayed –in order of increasing value.
Note:
Note that several data functions are involved in the entire data value chain: Data Integration to transform and transport the data, Metadata to maintain information on origin and history of the data, Data Security to protect the data, Data Quality to guarantee to right level of correctness and completeness and Data governance to define and guard the processes around data.
Abbreviations
The following abbreviations are used:
A next blog will elaborate on the model
Nice model, thanks for posting
I hear about this all the time! Great point of view on data value chain.
I agree with you that Business Rules Management should be a major capability in any enterprise. It is true BRM is not explicitly added to this model. However implicitly it is playing an important role. BRM might be part of CEP or BPM functionality (through a separate engine with a set of rules, or incorporated in the models), and in any enterprise application. Business Rules may even the result of discovery through analytics. The effect of business rules though is reflected in the model in meta-data (describing the origin and transformations of data) and data quality (results of proper rules lead to higher quality data).
Very useful. What in my view is missing from the model is the set of business rules. Intelligence not only comprises new information inferable from data (analytics) but also new information coming from policies, practices and knowledge in the organization that classify, judge and diagnose the data by the application of managed rules.